Professional Documents
Culture Documents
Lehnert 2008
Lehnert 2008
A REVISED
ELECTROMAGNETIC
THEORY WITH
FUNDAMENTAL
APPLICATIONS
EN REVIDERAD ELEKTROMAGNETISK
TEORI MED FUNDAMENTALA
TILLAMPNINGAR
SVENSKA
FYSIKARKIVET
Bo Lehnert
Royal Institute of Technology
Stockholm, Sweden
A Revised Electromagnetic
Theory with Fundamental
Applications
En reviderad elektromagnetisk teori
med fundamentala tillampningar
2008
Svenska fysikarkivet
Swedish physics archive
Svenska fysikarkivet (that means the Swedish physics archive) is a publisher registered with The National Library of Sweden, Stockholm. This book was typeset
using teTEX typesetting system and Kile, a TEX/LATEX editor for the KDE desktop.
Powered by Ubuntu Linux.
Edited by Dmitri Rabounski
c Bo Lehnert, 2008
Copyright
c Svenska fysikarkivet, 2008
Copyright
All rights reserved. Electronic copying and printing of this book for non-commercial,
academic or individual use can be made without permission or charge. Any part of
this book being cited or used howsoever in other publications must acknowledge this
publication.
No part of this book may be reproduced in any form whatsoever (including storage in
any media) for commercial use without the prior permission of the copyright holder.
Requests for permission to reproduce any part of this book for commercial use must
be addressed to the copyright holders.
Please send all comments on the book, requests for hardcopies, or requests to reprint
to the Author or the Publisher.
E-mail: Bo.Lehnert@ee.kth.se
E-mail: admin@ptep-online.com
Svenska fysikarkivet
ar ett f
orlagstryckeri registrerat hos Kungliga biblioteket. Denna
bok
ar typsatt med typs
attningssystemet teTEX och Kile, en TEX/LATEX-redaktion
f
or KDE-skrivbordsmilj
on. Utf
ord genom Ubuntu Linux.
Redakt
or: Dmitri Rabounski
c Bo Lehnert, 2008
Copyright
c Svenska fysikarkivet, 2008
Copyright
Eftertryck f
orbjudet. Elektronisk kopiering och eftertryckning av denna bok i ickekommersiellt, akademiskt, eller individuellt syfte
ar till
aten utan tillst
and eller kostnad. Vid citering eller anv
andning i annan publikation ska k
allan anges.
M
angfaldigande av inneh
allet, inklusive lagring i n
agon form, i kommersiellt syfte
ar f
orbjudet utan medgivande av upphovsr
attsinnehavaren. Beg
aran om tillst
and
att reproducera del av denna bok i kommersiellt syfte ska riktas till upphovsr
attsinnehavaren.
V
anligen skicka eventuella kommentarer till boken, efterfr
agan av tryckt kopia, eller
f
orfr
agan om nytryck till f
orfattaren eller f
orlaget.
E-post: Bo.Lehnert@ee.kth.se
E-post: admin@ptep-online.com
ISBN: 978-91-85917-00-6
Printed in Indonesia
Tryckt i Indonesien
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapter 1. Introduction
1.1 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
1.2 Some unsolved problems in conventional theory . . . . . . . . . . . . 12
1.3 The contents of the present treatise . . . . . . . . . . . . . . . . . . . . . . . . 14
Chapter 2. Short Review of Some Modified Theories
2.1 Theories based on additional vacuum currents. . . . . . . . . . . . . .15
2.1.1 Quantum mechanical theory of the electron . . . . . . . . . . . . . 16
2.1.2 Theory of the photon with a rest mass . . . . . . . . . . . . . . . . . . 16
2.1.3 Nonzero electric field divergence theory . . . . . . . . . . . . . . . . . 17
2.1.4 Nonzero electric conductivity theory . . . . . . . . . . . . . . . . . . . . 17
2.1.5 Single charge theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1.6 Related critical questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Theories based on further generalizations . . . . . . . . . . . . . . . . . . 19
2.2.1 Magnetic monopoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.2 Unification of electromagnetism and gravitation . . . . . . . . . 20
Chapter 3. Basis of Present Theory
3.1 Deduction of the space-charge current density . . . . . . . . . . . . . . 21
3.2 The extended field equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Comparison with the Dirac theory . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4 The quantum conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.5 The momentum and energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.6 The energy density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.6.1 Steady states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.6.2 Time-dependent states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Bo Lehnert
Contents
Bo Lehnert
Contents
To my wife
ANN-MARIE
Preface
Maxwells equations in the vacuum state have served as a guideline
and basis in the development of quantum electrodynamics (QED). As
pointed out by Feynman, however, there are important areas within
which conventional electromagnetic theory and its combination with
quantum mechanics does not provide fully adequate descriptions of
physical reality. These difficulties are not removed by and are not
directly associated with quantum mechanics. Instead electromagnetic
field theory is a far from completed area of research, and QED will
also become subject to the typical shortcomings of such a theory in its
conventional form. As a consequence, modified theories leading beyond
Maxwells equations have been elaborated by several authors. Among
these there is one approach which becomes the main subject of this treatise, as well as of its relation to other concepts. The underlying ideas
originate from some speculations in the late 1960s, and a revised theory
has then been gradually developed. An effort is made here to collect
the various parts of the theory into one systematic entity.
The present approach is based on a vacuum state which can become electrically polarized, such as to give rise to a local electric space
charge and a related nonzero electric field divergence. The latter can
as well be taken as a starting point of such a field theory. The condition of Lorentz invariance then leads to an additional space-charge
current which appears along with the conventional displacement current. Maxwells equations thus become a special case of the theory.
Relevant quantum conditions are imposed on the general solutions of
the field equations, to result in a first formulation of an extended quantum electrodynamical (EQED) approach. The nonzero electric field
divergence introduces an additional degree of freedom, thereby giving
rise to new features which are illustrated by several fundamental applications, such as those to steady electromagnetic states and additional
types of wave modes. These are represented by models of the electron,
neutrino and photon. They include solutions of so far not understood
problems, as represented by the radial force balance of the electron under the influence of its self-charge, the point-charge-like character of the
electron and the requirement of its finite self-energy, the nonzero angular momentum of the individual photon, and the wave-particle duality
and needle-radiation property of the latter.
10
Bo Lehnert
The author expresses his sincere thanks to Dr. Jan Scheffel for a
fruitful collaboration on a developed electron model. The author is
further indebted to to Prof. Kai Siegbahn for his shown interest and
encouragement concerning the present theory, and for many valuable
considerations, among these on the particle-wave dualism and the photoelectric effect. Also a number of interesting discussions with Prof. Nils
Abramson on photon physics and with Prof. Hans Wilhelmsson on the
concepts of field theory are gratefully acknowledged. The author is finally indebted to M. Sc. Anna Forsell and M. Sc. Kerstin Holmstrom for
valuable help with the manuscript of this treatise.
Stockholm, Sweden, 2007
Bo Lehnert
Chapter 1
INTRODUCTION
1.1
Background
12
Bo Lehnert
1.2
The failure of standard electromagnetic theory based on Maxwells equations appears in several important cases, such as those illustrated by the
following examples:
The absence of a steady electromagnetic equilibrium would cause
the electron to explode under the action of its electric eigencharge;
The electron behaves like a point charge, i.e. with an extremely
small radius, and it carries the smallest free elementary charge,
e. Standard theory is confronted with the infinite self-energy
of such a charge. A quantum electrodynamical treatment based
on a renormalization procedure has been applied to yield a finite
result, as deduced from the difference between two infinite forms.
Notwithstanding the success of this normalization, a physically
more satisfactory procedure is nevertheless needed concerning the
self-energy, as stated by Ryder21 ;
There also arises a so far unanswered question why the elementary
electronic charge has just the discrete experimentally determined
minimum value e, and how this becomes related to a possible
quantization process;
Light appears to be made of waves and simultaneously of particles. In conventional theory the individual photon is on one
hand conceived to be a massless particle, still having an angular
momentum (spin), and is on the other hand regarded as a wave
having the frequency and the energy h, whereas the angular
momentum is independent of the frequency. This dualism of the
wave and particle concepts is so far not fully understandable in
terms of conventional theory4 ;
Concerning the plane wave photon concept, which has no derivatives along its wave front, the energy becomes spread all over an
infinite volume. Such a concept cannot account for the needle-like
photon configuration with spatially concentrated energy which is
required for knocking out an atomic electron in the photoelectric effect. Nor can a plane wave produce the dot-shaped marks
which occur on the screen in double-slit experiments on individual
photon impacts22 ;
As shown by Heitler23 among others, the plane wave has no angular momentum and can therefore not represent the photon as a
boson particle. As a plane wave an individual photon of nonzero
Chapter 1 Introduction
13
14
Bo Lehnert
has a maximum for parallel polarization. At an arbitrary polarization angle the displacement does not acquire an intermediate
value, but splits into the two values of parallel and perpendicular polarization30. According to de Broglie and Vigier31 this behaviour cannot be explained by conventional theory;
In a rotating interferometer fringe shifts have been observed between light beams that propagate parallel and antiparallel with
the direction of rotation. Also this Sagnac effect requires an unconventional explanation according to Vigier32 ;
Electromagnetic wave phenomena and the related photon concept
remain somewhat of an enigma in more than one respect. The latter concept should in principle apply to wavelengths ranging from
about 1015 m of gamma radiation to about 105 m of long radio
waves. This includes the yet not fully understandable transition
from a low-density beam of individual photons to a nearly plane
electromagnetic wave of high photon density.
1.3
Chapter 2
1 2
2
A
A = 0 J ,
= 1, 2, 3, 4 ,
(2.1)
c2 t2
given in SI units. In this equation
i
,
A = A,
c
(2.2)
16
Bo Lehnert
where A and are the magnetic vector potential and the electrostatic
potential in three-space, and
J = (j, ic )
(2.3)
1
=0
c2 t
(2.4)
17
Vigier4 proposed a corresponding form of the four-current in the Procatype equation (2.1), as given by
2
1
2m0 c
i
J =
A,
.
(2.7)
0
h
c
The corresponding solution of the field equations was found to include longitudinal electric and magnetic field components, in addition
to the transverse ones. Thereby Evans42 used cyclically symmetric relations to derive a longitudinal magnetic field part, B(3), in the direction
of propagation.
2.1.3
C2 = c2 ,
(2.8)
18
Bo Lehnert
j = 0
+v E
(2.10)
t
in three-space. This approach includes longitudinal modes which do
not exist in Maxwellian theory. The same authors thereby conclude
that the Lienard-Wiechert potentials are incomplete, by not being able
to describe long-range instantaneous Coulomb interaction.
2.1.6
19
Two general objections may at a first glance be raised against the approaches which lead to a nonzero rest mass of the photon. The first
concerns the gauge invariance. A positive answer to this point has recently been given by an analysis in terms of covariant derivatives56. This
is consistent with an earlier conclusion that gauge invariance does not
require the photon rest mass to be zero57. Concerning the present theory
by the author, the deductions in the following Chapter 3 also show that
the field equations are gauge invariant, even when they lead to a nonzero
photon rest mass. This invariance applies to all systems where the electric and magnetic fields are unaffected by changes in the corresponding
potentials, when these are subjected to a gauge transformation.
The second objection concerns the supposition that a nonzero rest
mass provides a photon gas with three degrees of freedom, i.e. two
transverse and one longitudinal. This would alter Plancks radiation
law by a factor of 23 , in contradiction with experience4,9. There are,
however, arguments which resolve also this problem. A detailed analysis by Evans and Vigier4 shows that the longitudinal magnetic field
component, which is associated with the spin and the rest mass, cannot
be involved in the process of light absorption. A further argument is
due to the fact that transverse photons are not able to penetrate the
walls of a cavity, whereas this is the case for longitudinal photons which
would then not contribute to the thermal equilibrium41.
In this connection it should finally be mentioned that the equations
of state of a photon gas have been treated by Mezaros58 and Molnar59
who find that Plancks distribution cannot be invariant to adiabatic
changes occurring in an ensemble of photons. This dilemma is due to the
fact that imposed changes cannot become adiabatic and isothermal at
the same time. Probably this contradiction may be resolved by including
a longitudinal part into the magnetic flux, but further analysis appears
to become necessary.
2.2
There also exist approaches of a more general character than those based
only on additional vacuum currents, as given by the following two concepts.
2.2.1
Magnetic monopoles
20
Bo Lehnert
Chapter 3
The current density (2.3) appearing in the right-hand side of the Procatype equation (2.1) is here required to transform as a four-vector. This
implies that the square of J should become invariant to a transition
from one inertial frame K to another such frame K 0 . Then Eq. (2.3)
yields
j2 c2 2 = j02 c2 02 = const.
(3.1)
In addition, the corresponding three-space current density j is required to exist only when there is also an electric charge density associated with the nonzero electric field divergence. This implies that the
constant in Eq. (3.1) vanishes. Since J has to become a four-vector, j
and must behave as space and time parts of the same vector. Therefore the choice of a vanishing constant in Eq. (3.1) becomes analogous
to the choice of the origin at x = y = z = t = 0 in the Lorentz invariant
relation
(3.2)
x2 + y 2 + z 2 c2 t2 = 0
22
Bo Lehnert
between the rectangular coordinates (x, y, z) and time t for a propagating light wave. Consequently the three-space current density becomes
j = C = 0 (div E)C ,
C2 = c2 ,
(3.3)
(3.4)
(3.5)
div E =
,
0
(3.6)
(3.7)
(3.8)
(3.9)
(3.10)
23
where the first term of the right-hand member of Eq. (3.6) and Eq. (3.10)
are the new parts being introduced by the present theory. The nonzero
electric field divergence introduces an asymmetry in the way which the
electric and magnetic fields appear in these equations.
The presence in Eqs. (3.10) and (3.6) of the dielectric constant 0 and
the magnetic permeability 0 of the conventional vacuum state may require further explanation. Liquid and solid matter consist of atoms and
molecules which often behave as electric and magnetic dipoles. Such
matter can thus become electrically polarized and magnetized when
external electric and magnetic fields are being imposed. Then the constants 0 and 0 have to be replaced by modified values and , which
take these polarization effects into account. One consequence of this is
that a conventional plane
electromagnetic wave would propagate
1 1/2
in such
1/2
matter at a velocity
being smaller than the velocity c = 010
in empty vacuum space. In the present extended theory on the vacuum state, however, no electrically polarized and magnetized atoms or
molecules are present. There are only electromagnetic fluctuations due
to the zero point field as revealed by the Casimir effect, and there can
also arise electric space charges out of the vacuum, such as in the case
of positron-electron pair formation. In other words, the vacuum state
is here conceived as a background of empty space upon which are superimposed zero point electromagnetic wave-like fluctuations, as well as
phenomena of a more regular character which could take the form of the
wave phenomena and field configurations to be treated in the present
theory. In this respect Eq. (3.10) becomes identical with that used in
the conventional theory on hot plasmas which contain freely moving
charged particles of both polarities, in a background of empty vacuum
space and where electric charge separation and a resulting electric field
can arise, such as in the case of plasma oscillations. There may exist an
interaction between the zero point field and regular modes, not being
treated here at this stage.
As mentioned earlier in 2.2.1, the question may also be raised why
only div E and not div B should be permitted to become nonzero in an
extended approach. It has then to be noticed that the nonzero electric
field divergence has its support in the observed pair formation and the
vacuum fluctuations, whereas a nonzero magnetic field divergence is so
far a possible but not proved theoretical supposition. With Dirac63
we shall therefore leave the magnetic monopole concept as an open
question, but not include it at the present stage in the theory.
As shown later, this theory will lead to a small but nonzero photon
rest mass. This raises a further question about the gauge invariance al-
24
Bo Lehnert
A0 = A + 5 ,
0 =
.
(3.11)
t
When being inserted into Eqs. (3.8) and (3.9), these gauge relations
lead to field strengths E0 = E and B0 = B which thus become invariant.
The field equations (3.6) and (3.7) only contain the field strengths and
therefore become gauge invariant.
As discussed later in 11.4, it has during later years been realized
that the electromagnetic four-potential (2.2) does not only serve the
purpose of a mathematical intermediary, but also can have a physical
meaning of its own. To the discussion on the gauge invariance it has
therefore to be added that there are situations where the choice of gauge
has indeed a physical influence, such as in the case of a curl-free magnetic
vector potential. We return to this question in Chapter 11.
The basic equations (3.3)(3.10) can as well be derived from the
more condensed form of a Lagrangian density
1
L = 0 (E2 c2 B2 ) + j A
(3.12)
2
given e.g. by Goldstein70. This constitutes a formally more elegant
representation, but it does not lead to more physical information than
what is already contained in the original basic equations.
The comparatively simple formalism of the present theory relates it
in a surveyable way to a number of physical concepts and applications.
This can be taken as an advantage, in particular as there are examples
to be shown later where the obtained results seem to agree rather well
with experimental experience.
3.3
25
U
0
= u(x, y, z)
(3.13)
U ,
0
where u is an arbitrary function and U a constant. This yields a charge
density
Uu
= 2e U
u
(3.14)
and the corresponding current density components
jz = c ,
jx = 0 , and jy = 0 ,
(3.15)
where a bar over U and u indicates the complex conjugate value. Other
analogous forms can be chosen where instead jy =c or jx = c .
This result and expressions (2.5) and (2.6) indicate that there is
a connection between the present theory and that by Dirac. But the
former theory sometimes applies to a larger class of phenomena, in the
capacity of both bound and of free states. Thereby the elementary
charge is a given quantity in Diracs theory, whereas the total net charge
of the present approach is deduced from the field equations in a steady
axisymmetric state, as shown in the following Chapters 5 and 6.
3.4
26
Bo Lehnert
(E B) ,
t
(3.16)
2
S n dS = Fe + Fm +
g dV .
(3.17)
t
Here dS and dV are surface and volume elements, n is the surface
normal,
Z
Z
Fe = E dV
Fm = C B dV
(3.18)
are the integrated electric and magnetic volume forces, and
g = 0 E B =
1
S
c2
(3.19)
27
the magnitude and direction of the energy flow in space. The component
Sjk of the stress tensor is the momentum that in unit time crosses a unit
element of surface in the j direction, and the normal of which is oriented
along the k axis. This implies that Eq. (3.17) represents Newtons third
law of the electromagnetic field, where the right-hand member stands
for the rate of change of the total momentum. The latter consists of
one contribution due to the mechanical volume force (3.18), and one
due to the momentum (3.19) of the electromagnetic field. The results
(3.16) and (3.17) differ from the conventional ones in the appearance of
terms which include a nonzero charge density , and that there is both
an electric and a magnetic volume force in the vacuum state.
Scalar multiplication of Eq. (3.6) by E and Eq. (3.7) by B0 , and a
subtraction of the resulting expressions, yields the local energy equation
1
1
div S =
div (E B) = E C + 0
(E2 + c2 B2 ) . (3.20)
0
2 t
This equation differs from the conventional Poynting theorem
through the appearance of the term E C. That such a difference
arises in some of the modified theories has also been emphasized by
Evans et al.6 and by Chubykalo and Smirnov-Rueda53. These investigators note that the Poynting vector in the vacuum is only defined in
terms of transverse plane waves, that the case of a longitudinal magnetic field component leads to a new form of the Poynting theorem, and
that the Poynting vector can be associated with the free electromagnetic field only. We shall later return to this question when considering
axisymmetric wave packets.
The integral form of expression (3.20) becomes
Z
Z
Z
1
S n dS + E C dV = 0
(E2 + c2 B2 ) dV.
(3.21)
2
t
The customary interpretation of this relation is as follows24 . It is
assumed that the formal expression for the energy density stored in the
electromagnetic field is the same as in a stationary state. Then the righthand side of (3.21) represents the rate of decrease of the electromagnetic
energy. This loss is accounted for by an input of energy due to E C,
plus an energy outflow in the direction of the Poynting vector S.
This interpretation of Poyntings theorem is, however, open to some
criticism24. First, from a volume integral representing the total energy
of the field no rigorous conclusion can be drawn with regard to its local
distribution. Second, the question may be raised as to the propriety
of assuming that the energy density and its rate of change remain the
28
Bo Lehnert
same for rapid as well as for quasi-steady changes. Third, even though
the total flow of energy through a closed surface may be represented
correctly by the Poynting vector, one cannot conclude definitely that
the local intensity of energy flow at a local point is given by the same
vector. The classical interpretation of Poyntings theorem thus appears
to rest to a considerable degree of hypothesis. It should therefore not
be considered as an absolute truth, but as a self-consistent and useful
analytical formulation24.
It has to be remembered that relations (3.16) and (3.20) have merely
been obtained from a rearrangement of the basic equations. They therefore have the form of identities by which equivalent expressions are obtained for the momentum and energy from the stress tensor.
3.6
A
A
= 2 + div A + +
. (3.23)
div +
t
t
t
Using Eqs. (2.1)(2.4) and (3.3)(3.10), the sum of expression (3.23)
and expression (3.22) multiplied by c2 leads after some deductions to
wf ws = wfs + div F ,
(3.24)
where
1
wf =
2
B2
2
0 E +
,
0
1
1
( + j A) = ( + C A) ,
2
2
"
#
2
1
A
2A
A
wfs = 0
A
+
A
,
2
t
t
t
t2
A
2
F = +
+ c2 A curlA = E + c2 A B .
0
t
ws =
(3.25)
(3.26)
(3.27)
(3.28)
Here wf can be interpreted as a field energy density being associated with the field strengths E and B, and ws as a source energy
29
Steady states
In a time-independent state where wfs = 0 the electrostatic and magnetostatic energy densities have been deduced in an unquestionable way
from the work exerted on the charges and currents24 . This leads to
the expression (3.26) for the source energy density. When applying this
result care is necessary to secure a physically correct choice of the reference (zero) levels of the potentials A and . It has also to be observed
that the relation
Z
Z
wf dV = ws dV
(3.30)
only holds when the surface integral of Eq. (3.29) vanishes, and this is
not always the case.
Here wf > 0 according to its quadratic form (3.25), whereas ws can
consist both of positive and of negative local contributions. This implies
that even negative energy states cannot be excluded for certain when
div F in Eq. (3.24) differs from zero.
Since wf and ws represent different spatial distributions of energy,
we also have the inequality
Z
Z
f wf dV 6= f ws dV
(3.31)
for the moment of the energy density being formed with any spatial
function f 6= const.
In a steady state, and in absence of surface charges and currents, it
is incontrovertible that the density (3.26) represents the local work done
on the charges and currents, as obtained from the primary deductions
of the energy density24 . This holds also when the electromagnetic field
becomes divergent at a particular point, such as at the origin of the
spherical coordinates in an axially symmetric configuration. It is then
30
Bo Lehnert
only necessary to choose a physically correct zero level of the electrostatic potential . Here we consider the corresponding total energy
Z
(3.32)
W = ws dV
which is finite and applies to configurations generated by charges and
currents that are restricted to a limited region of space. For the boundaries of the integral (3.32) there are two possibilities to be taken into
account:
When the density ws is convergent everywhere and vanishes rapidly at large distances from the origin, the integral (3.32) can be
extended throughout space;
When the density ws still vanishes at large distances, but becomes
divergent at the origin, an inner surface Si is defined which encloses the origin and forms a lower limit of the integral (3.32). The
divergence of the energy density can then be handled in terms of
the integral as it stands, and by applying a special analytic procedure based on a smallness parameter which defines a shrinking
radius of the inner surface Si . This scheme will be demonstrated
later in Chapters 5 to 7.
Here one may ask the question if it would become convenient to
divide the electromagnetic field into one source part (As , s ) where
curlcurl As 6= 0 and 2 s 6= 0, and one sourceless part (Av , v ) where
curlcurl Av = 0 and 2 v = 0. This does, however, not seem to simplify
the analysis which includes all features of the field. The following points
have namely to be observed:
For any given form of the electromagnetic potentials A and the
field strengths E and B as well as the source energy density ws
become uniquely determined;
In an axially symmetric geometry with a space charge density
being mainly concentrated to a region near the origin of a spherical
frame of reference, these potentials can in principle be chosen such
as to approach the limit determined by the sourceless part at
large radial distances;
In such a geometry where the dominating contributions to the
source energy density originate from regions near the origin, the
integrated total energy (3.32) will become independent of the special asymptotic forms of A and at large radial distances. Examples of this will be given by the electron and neutrino models
of Chapters 6 and 7.
3.6.2
31
Time-dependent states
(3.33)
(3.34)
in addition to the basic equations (3.6) and (3.7). The result is then an
overdetermined and unacceptable system of nine equations for the six
components of E and B.
With respect to the volume forces, the present method of proceeding
has therefore been based on an analysis of the resulting integrated force
balance in every particular case. This will later be demonstrated in 6.8
for an electron model, in 7.4 for a neutrino model, and in 9.3.3 for a
photon model.
32
Bo Lehnert
3.8
Chapter 4
Figure 4.1: New features introduced by the present theory. The lowest row
represents possible areas of application.
Steady phenomena
In a time-independent case the space-charge current density makes possible the existence of steady electromagnetic states which are absent
34
Bo Lehnert
C
0
(4.1)
Time-dependent phenomena
2 2
2
c E+ c +C
(div E) = 0
t2
t
(4.2)
for the electric field. From the corresponding solution the magnetic
field B can be obtained by means of Eq. (3.7). For the magnetic field B
a general differential equation is on the other hand not available from
the basic equations, as long as div E remains nonzero. A divergence
operation on Eq. (3.6) further yields
+ C (div E) = 0
(4.3)
t
in combination with Eq. (3.5). In some cases this equation becomes
useful to the analysis, but it does not introduce more information than
35
that already contained in Eq. (4.2). The result (4.3) is directly related
to the condition of charge conservation
div j =
(4.4)
Chapter 5
(r0 )2
= D = D + (sin )2 (CA) ,
(5.1)
0
where the dimensionless radial coordinate
r
=
(5.2)
r0
has been introduced, r0 is a characteristic radial dimension, and the
operator D is given by
2
cos
D = D + D , D =
2
, D = 2
. (5.3)
sin
For string-shaped states a frame (r, , z) of cylindrical coordinates
is instead adopted, with both and z as ignorable quantities, and with
z directed along the axis of the configuration. Then Eqs. (4.1) reduce to
(r0 )2
1
(5.4)
= D = D + 2 (CA)
0
1 d
d
D=
.
(5.5)
d
d
5.1
(5.6)
37
= 1 + (sin2 ) D F ,
0
= 2 D 1 + (sin2 ) D F ,
r 0 2
(5.7)
(5.8)
(5.9)
(5.10)
results in
CA = 2 DF ,
= 1 + 2 D F ,
0
= 2 D 1 + 2 D F .
2
r0
(5.11)
(5.12)
(5.13)
Particle-shaped states
f (, ) = (sin ) D 1 + (sin2 ) D G ,
g(, ) = 1 + 2(sin2 ) D G
(5.14)
(5.15)
38
Bo Lehnert
Iq = f ,
M0 = 0 C r02 G0 JM , IM = (sin )f ,
0
r0 G20 Jm ,
c2
0 C 2 2
s0 =
r0 G0 Js ,
c2
m0 =
(5.17)
Im = f g ,
(5.18)
Is = (sin )f g .
(5.19)
(5.16)
(5.20)
39
and (3.16) and investigated later in detail in 6.8. This leads under certain conditions to a self-confined balance between the electromagnetic
forces. The electron can then be prevented from exploding under
the action of its electric eigencharge, as otherwise being predicted by
conventional theory24,25.
At this point a further step is taken by imposing the restriction of a
separable generating function
G(, ) = R() T () .
(5.21)
(5.22)
IM = (sin ) Iq ,
(5.23)
Im = 0 3 R2 + (0 4 + 1 3 ) R (D R) + 1 4 (D R)2 +
+ 2 3 RD (D R) + 2 4 (D R)[D (D R)] ,
Is = (sin ) Im ,
(5.24)
(5.25)
where
(5.26)
(5.27)
2 = (sin3 ) T ,
(5.28)
3 = T 2 (sin2 )(D T ) ,
(5.29)
4 = 2 (sin2 ) T .
(5.30)
Among the possible forms to be adopted for the radial function R()
and the polar function T (), we will here consider the following cases:
The radial part R can become convergent or divergent at the origin
= 0 but it always goes strongly towards zero at large distances
from it;
The polar part T is always finite and has finite derivatives. It can
become symmetric or antisymmetric with respect to the equatorial plane (mid-plane) defined by = 2 .
The restriction (5.21) of separability becomes useful when treating
configurations where the sources and j and the corresponding energy
density are mainly localized to a region near the origin, such as for a particle of limited extent. Then the far-field properties, at great distances
40
Bo Lehnert
from the origin, have a negligible influence on the integrated field quantities (5.16)(5.19). This restriction is also supported indirectly by the
results which follow and which in several respects seem to be consistent
with experimental facts.
5.2.1
2 dR
2
2 d
2 dR
+ 2
d ,
d2
d
0
(5.31)
d
dT
2
Jq = (sin )
(5.33)
(sin )(D T ) + (sin )
.
d
d 0
For all functions T with finite derivatives at = (0, ), it is then
seen that Jq and q0 vanish in general.
Turning next to the magnetic moment (5.17) and the integral JM ,
it is first observed that
Z
Z
Z
1
1
R d =
(5.34)
(D R) d =
D (D R) d
2 0
4 0
0
as obtained by partial integration. Using this relation, Eq. (5.23) yields
Z
Z
JM =
R d
(sin )(0 21 + 42 ) d JM JM . (5.35)
0
d
3
JM = (sin )
(sin )(D T 2T )
d
0
(5.36)
41
42
Bo Lehnert
5.2.4
43
String-shaped states
(5.37)
Here the exponential factor secures the convergence at large distances from the axis of symmetry where H() is finite.
5.3.1
In analogy with Eq. (5.16) the net electric charge per unit length becomes
Z
q0
= 20 G0 Jq ,
Jq =
fq d ,
(5.38)
Ls
0
where Ls is the length of the string and
fq = D(1 + 2 D) G =
r02
0 G0
(5.39)
according to Eq. (5.13). With the operator (5.5) the integral (5.38)
becomes
d
d
dG
Jq =
G
.
(5.40)
d
d
d
0
The generating function (5.37) then yields
n
Jq = 2 e [(3 )(H 0 H) + (3 2)(H 00 H 0 ) +
o
+ (H 000 H 00 )]
,
(5.41)
Z
C
fq
b() ,
b() =
d .
B() = G0
(5.42)
c2 r0
44
Bo Lehnert
b() = e 2 + 4 2 .
(5.43)
also vanishes.
5.3.3
0 C 2 2
h
r G Js =
.
c2 0 0
4
(5.45)
45
!1/2
q
f0 Jq2
2 0 ch
0
q =
,
(5.46)
,
f0 =
e
2Js
e2
where q is a dimensionless charge being normalized with respect to
the experimentally determined electronic charge e, and f0 137.036
is the inverted value of the fine-structure constant.
5.4.2
According to Schwinger and Feynman82, and Dirac37 the quantum condition on the magnetic moment of a charged particle such as the electron
becomes
M 0 m0
1
(5.47)
= 1 + M ,
M =
q 0 s0
2f0
which shows excellent agreement with experiments. Here the unity term
of the right-hand member is due to Dirac who obtained the correct
Lande factor by considering the electron to be situated in an imposed
external magnetic field. This leads to a magnetic moment being twice
as large as that expected from an elementary proportionality relation
between the electron spin and the magnetic moment21. Further, in
Eq. (5.47) the term M is a small quantum mechanical correction due to
Schwinger and Feynman, as obtained from an advanced analysis where
the electron is considered to emit and absorb a photon during its way
from one place to another.
Conditions (5.45) and (5.47) can also be made plausible by elementary physical arguments based on the present picture of a particleshaped state of self-confined radiation. In the latter picture there
is a circulation of radiation at the velocity of light around the axis of
symmetry. If this circulation takes place at the average characteristic
radial distance r0 , the corresponding frequency of revolution would bec
come ' 2r
. In combination with the energy relations W = m0 c2 and
0
W = h by Einstein and Planck, one then obtains
r0 m0 '
h
,
2c
(5.48)
where r0 has the same form as the Compton radius. With the total
46
Bo Lehnert
1
q0 h
q0 c r0 =
,
2
4 m0
(5.49)
Chapter 6
n
X
=1
> 0,
(6.1)
(6.2)
48
Bo Lehnert
sin3 cosn3 +
3
n
+
sin5 cosn5 . . . (6.3)
5
n
n
2
n
n2
cos n = cos
sin cos
+
sin4 cosn4 . . . (6.4)
2
4
A separable generating function (5.21) is used in this analysis which
in the first place aims at the region near the origin where the charge density is mainly being concentrated. As will be shown later, and pointed
out in 3.6.1, the dominating contributions to the source energy density then originate from this region. The final result obtained from the
corresponding integrals is independent of the special asymptotic forms
of the electromagnetic potentials at large radial distances. To achieve
0 0
(6.5)
This would, however, much complicate the analysis, without changing the final results which are of main importance to this investigation.
6.2
The radial form given by Eq. (6.1) can now be inserted into the integrands (5.22)(5.25). The expressions for D R and D (D R) give rise
to sets of terms with different negative powers of . Since the radial integrals (5.20) are extended to very low limits k , they become dominated
by contributions from terms of the strongest negative power. Keeping
these contributions only, the integrands reduce to
Ik = Ik Ik ,
where
and
Iq = R ,
IM = R ,
k = q, M, m, s,
Im = R2 ,
Is = R2 ,
(6.6)
(6.7)
Iq = 0 ( 1) 1 + 2 ( 1)2 2 ,
(6.8)
IM = (sin ) Iq ,
(6.9)
Im = 0 3 ( 1)(0 4 + 1 3 ) +
+ 2 ( 1)2 (1 4 + 2 3 ) 3 ( 1)3 2 4 ,
(6.10)
Is = (sin ) Im .
(6.11)
49
(6.12)
where
Jq =
Jm =
1
(1) ,
1 q
JM =
1
(21) ,
2 1 m
Js =
and
1
(2)
,
2 M
(6.13)
1
2(1) ,
2( 1) s
(6.14)
Jk =
Ik d .
(6.15)
c0 > 0 ,
0 < 1.
(6.16)
Jq
( 1)
1
q
20 C
c30 G30
JM Jm
(6.17)
3
21
( 2)(2 1) 2
M m
2
0 C 2 2 Js
s0 =
c0 G0
.
c2
2( 1) 1
s
c2
(6.18)
(6.19)
50
Bo Lehnert
(6.20)
and that the radial parameter has to approach the value 2 from
above, i.e.
( 1) = 2 + ,
0 6 1 ,
2+ .
(6.21)
3
For the quantities M0 and m0 to become finite separately when
approaches zero, the corresponding conditions become less straightforward14, but M0 m0 can be made to follow the scaling of Eq. (6.20)
also in such a case.
From the earlier obtained results (5.33) and (5.36) it is further seen
that the contribution from 0 in Eq. (6.8) vanishes as well as JM when
= 2. In the limit (6.21) the integrands (6.8)(6.11) can then be replaced by
Iq = 21 + 42 ,
(6.22)
IM
= (sin )(1 + 42 ) ,
Im = 0 3 2(0 4 + 1 3 ) + 4(1 4 + 2 3 ) 8 2 4 ,
Is = (sin ) Im .
(6.23)
(6.24)
(6.25)
AM
q0 = 20 c0 G0 Aq /3 ,
20 C 3 3
c G AM Am ,
c2 0 0
1 0 C
(6.26)
(6.27)
(6.28)
(6.29)
51
6.3
Magnetic flux
G0
(sin3 ) DG
c
(6.30)
which vanishes at = (0, ). Making use of Eqs. (5.3), (6.1), and (6.16)
the flux becomes
= 2
c0 G0
sin3 ( 1) + 2( 1) + 2 T
C
D T 1 e .
(6.31)
where
c0 G0 /3
0 = = , =
= 2
A ,
2
C
A = D T 2T =
(6.32)
(6.33)
for = 2. Also here the result can be considered to include the ampli
tude G0 /3 . The flux (6.32) can be regarded as being generated by a
configuration of thin current loops. These have almost all their currents
located to a spherical surface with the radius = , and the corresponding magnetic field lines cut the equatorial plane at right angles.
For C < 0 and q0 < 0 there is then an upward flux 0 within the
inner region 0 < < of the equatorial plane, being equal to a downward flux 0 within the outer region > of the same plane.
52
Bo Lehnert
Figure 6.1: Crude outline of a magnetic field configuration in the case of one
magnetic island situated above and one below the equatorial plane = 2 .
Only the upper half-plane is shown, and the figure is axially (rotationally)
symmetric around the vertical axis = 0. In the interval 1 6 6 2 there is
a magnetic flux into the spherical surface = (dashed circular line) being
equal to the inward magnetic island flux only. The outward flux parts in the
intervals 0 6 6 1 and 2 6 6 2 consist of the outward main flux, plus an
outward island flux.
It has to be observed that the flux (6.32) is not necessarily the total flux which is generated by the current system as a whole. There
are cases in which magnetic islands are formed above and below the
equatorial plane, and where these islands possess an isolated circulating extra flux which does not intersect the same plane. The total flux
tot then consists of the main flux 0 of Eq. (6.32), plus the extra
island flux i , which can be deduced from the function (6.31), as will
be shown later. This type of field geometry turns out to prevail in the
analysis which follows, and it leads to a magnetic configuration being
similar to that outlined in Fig. 6.1.
In the analysis of the contribution from the magnetic islands we
introduce the normalized flux function defined by Eq. (6.31) in the upper
half-plane of the sphere = 1. It becomes
( = , ) C /3
(6.34)
= (sin3 )(D T 2T ) .
c0 G0
2
53
Within the region near this spherical surface the function (6.31) and
the corresponding vector potential both consist of the product of two
separate functions of and . This also applies to the components of the
magnetic field given by Eq. (3.8). The polar component B is then seen
to be proportional to the
function of Eq. (6.34),
whereas the radial component B becomes proportional to d
. In all cases
d
to be treated here the situation outlined in Fig. 6.1
is then represented by a
magnetic near-field configuration being demonstrated in Fig. 6.2.
In the range of increasing , the component B
then vanishes along the
lines = 1 and = 2 ,
Figure 6.2: Outline of the magnetic field con- whereas the component
figuration corresponding to the present de- B vanishes along the
tailed deductions. The figure shows the field in lines = and = .
3
4
the region near the spherical surface = . The Consequently, the increaradial magnetic field component B vanishes
se of from the axis at
along the lines = 1 and = 2 , whereas the
= 0 first leads to an inpolar component B vanishes along the lines
creasing flux up to a
= 3 and = 4 .
maximum at the angle
= 1 . Then there follows an interval 1 < < 2 of decreasing flux,
down to a minimum at = 2 . In the range 2 6 6 2 there is again
an increasing flux, up to the value
0 =
= A ,
(6.35)
2
which is equal to the main flux.
We further introduce the parts of the integrated magnetic flux
defined by
(2 ) ,
1 = (1 ) (0) = (1 ) ,
2 =
(6.36)
2
and associated with the ranges 0 < < 1 and 2 < < 2 of Figs.6.1
and 6.2. The sum of these parts includes the main flux 0 , plus an
outward directed flux from one magnetic island. The contribution from
54
Bo Lehnert
(6.37)
ff =
2(1 + 2 ) 0
,
0
(6.38)
Quantum conditions
!1/2
f0 A2q
q =
(6.39)
As
according to Eqs. (6.12)(6.15), (6.20) and (6.26) in the limit = 2.
For the magnetic moment condition (5.47) reduces to
AM Am
= 1 + M
Aq As
(6.40)
when applying Eqs. (6.27)(6.29). The condition (6.40), and its counterpart of Eq. (5.47), has earlier been made plausible by the simple
physical arguments of 5.4.2. This also seems to be supported by the
forms of the integrands (6.8)(6.11) according to which the local relation IM Im = Iq Is is obtained. However, as has become obvious from
the present analysis and Eqs. (6.21)(6.23), the detailed deductions of
the electron model do not become quite as simple.
Magnetic flux quantization is expressed by condition (5.50). Combination of Eqs. (6.35), (6.17), (6.19) and (6.26) then yields
8fq A Aq = As ,
(6.41)
(6.42)
6.5
55
The previous analysis has shown that well-defined convergent and nonzero integrated physical quantities can be obtained in a point-charge-like
steady state, thereby avoiding the problem of an infinite self-energy. Attention may here be called to Ryder21 who has stressed that, despite the
success of the conventional renormalization procedure, a more physically
satisfactory way is needed concerning the infinite self-energy problem.
Possibly the present theory could provide such an alternative, by tackling this problem in a more surveyable manner. The finite result due
to a difference between two infinities in renormalization theory, i.e.
by adding extra counter-terms to the Lagrangian, is then replaced by a
finite result obtained from the product of an infinity with a zero,
as determined by the combination of the present divergent integrands
with a shrinking characteristic radius.
As expressed by Eq. (6.16), the latter concept also has an impact on
the question of Lorentz invariance of the electron radius. In the limit
= 0 of a vanishing radius r0 corresponding to a structureless mass
point, the deductions of this chapter will thus in a formal way satisfy
the requirement of such an invariance. At the same time the obtained
solutions can also be applied to the physically relevant case of a very
small but nonzero radius of a configuration having an internal structure.
6.6
The elementary electronic charge has so far been considered as an independent and fundamental physical constant of nature, being determined by measurements only84. However, since it appears to represent
the smallest quantum of free electric charge, the question can be raised
whether there is a more profound reason for such a minimum charge to
exist, possibly in terms of a quantized variational analysis.
The present theory can provide the basis for such an analysis. In a
first attempt one would use a conventional procedure including Lagrange
multipliers in searching for an extremum of the normalized charge q of
Eq. (6.39), under the subsidiary quantum conditions (6.40)(6.42) and
(6.38). The available variables are then the amplitudes a21 and a2 of
the polar function (6.2). Such an analysis has unfortunately been found
to become quite complicated83, partly due to its nonlinear character
and a high degree of the resulting equations. But there is even a more
serious difficulty which upsets such a conventional procedure. The latter
namely applies only when there are well-defined and localized points of
an extremum, in the form of a maximum or minimum or a saddle-point
56
Bo Lehnert
in parameter space, but not when such single points are replaced by
a flat plateau which has the effect of an infinite number of extremum
points being distributed over the same space.
A plateau-like behaviour is in fact what occurs here83, and an alternative approach therefore has to be elaborated. The analysis then
proceeds by successively including an increasing number of amplitudes
(a1 , a2 , a3 , . . .) which are being swept (scanned) across their entire
range of variation. The flux factor (6.42) has at the same time to
be determined in a self-consistent way through an iteration process.
For each iteration the lowest occurring value of q can then be determined. Thereby both conditions (6.40) and (6.41) and the flux factors
of Eqs. (6.38) and (6.41) include variable parameters. Thus the flux
factor ff of Eq. (6.41) does not become constant but varies with the
amplitudes of the polar function (6.2) when there are magnetic islands
which contribute to the magnetic flux. At a first sight this appears
also to result in a complicated and work-consuming process. However,
as demonstrated later in 6.6.2 and 6.6.3 on a flat plateau behaviour,
this simplifies the corresponding iteration scheme and the physical interpretation of its results.
In the numerical analysis which follows, solutions with two real roots
have always been found. Of these roots only that resulting in the lowest
value of q will be treated in detail in the following parts of this chapter.
6.6.1
As a first step the solutions for only two nonzero amplitudes, a1 and
a2 , are considered as obtained from conditions (6.40)(6.42). Matching
of the two flux factors of Eqs. (6.38) and (6.47) is then illustrated by
Fig. 6.3 where the self-consistent value of the normalized charge becomes
q ' 1.01. This value will, however, not turn out to be the lowest one
obtained by means of the present variational analysis on a larger number
of nonzero amplitudes.
In the next step four amplitudes (a1 , a2 , a3 , a4 ) are included in the
polar function (6.2). The normalized charge q can then be plotted
as a function of the two amplitudes a3 and a4 which are swept across
their entire range of variation, but for a fixed flux factor f = fq = 1.82.
The result is shown in Fig. 6.4 for the lowest of the two roots obtained
for q . In fact, the corresponding result of a self-consistent analysis
which takes Eq. (6.42) into account, leads only to small modifications
which hardly become visible on the scale of Fig. 6.4. From the figure is
seen that there is a steep barrier in its upper part, from which q drops
57
normalized charge q obtained through match- found to become plateauing of the flux factors fq and ff in the case like. The details of these
of only two nonzero amplitudes a1 and a2 of deductions will be discussed in the coming subsecthe polar function.
tions 6.6.2 and 6.6.3.
The multistep iteration process which leads to a self-consistent flux
factor for the two independent relations (6.38) and (6.41) requires a
rather high degree of accuracy, also being related to a determination of
the zero points 1 and 2 of the derivative d
in Figs. 6.2 and 6.5. This
d
problem can be tackled more efficiently in terms of a special analysis
restricted to the plateau regime, as shown in the next subsection.
6.6.2
The plateau behaviour demonstrated by Fig. 6.4 and obtained from the
numerical analysis can be taken as a basis for an asymptotic theory,
being adapted to the limit of large moduli of the amplitudes in the
polar function (6.2), i.e. for the behaviour at large distances from the
origin a3 = a4 = 0 in Fig. 6.4. To include all points of large amplitudes
which are contained within the plateau region far from the origin, we
now restrict the range of the amplitudes through the condition
ai
= a +
ki
i = 3, 4, . . .
(6.43)
Here ki are scaling factors associated with each amplitude, and which
in principle could become functions of the parameters which define the
plateau region. Since the amplitudes a1 and a2 cannot vanish in this
58
Bo Lehnert
Figure 6.4: The normalized electron charge q |q0 /e| as a function of the
two amplitudes a3 and a4 , for solutions satisfying the subsidiary quantum
conditions for a fixed flux factor f = fq = 1.82, and being based on a polar
function T having four amplitudes (a1 , a2 , a3 , a4 ). The profile of q consists
of a steep barrier in the upper part of the figure, and a flat plateau in
the lower part. The plateau is close to the level q = 1. The figure only
demonstrates the ranges of the real solutions of the first (lowest) root. The
deviations of this profile from that obtained for the self-consistent solutions
which obey condition (6.42) are hardly visible on the scale of the figure.
(6.44)
where
Tn =
n
X
(6.45)
(6.46)
=3
59
Figure 6.5: The normalized magnetic flux d/d per unit angle at the
spherical surface = . The sum of the areas 1 and 2 represents the total
outflux, and the area i = 1 + 2 0 is the counter-directed flux due to
one magnetic island.
of Fig. 6.4 the analysis is then performed along lines within the plateau
which at different angles originate from the point a3 = a4 = 0. With the
asymptotic forms (6.43)(6.46) inserted into expressions (5.26)(5.30)
and (6.34) the corresponding field quantities become
= (T) ,
a
Aq
AM
Aq =
,
AM =
,
a
a
As
Am
As = 2 ,
Am = 2 ,
a
a
and
A(T )
A =
= A(T) ,
a
= (T ) = (T) ,
1 +
2)
0
2(
.
ff =
0
(6.47)
(6.48)
(6.49)
(6.50)
(6.51)
(6.52)
60
Bo Lehnert
forms
q =
f0 A2q
As
!1/2
1/2
f0 Aq
,
8 fq A
(6.53)
AM Am
= 1 + M ,
Aq As
(6.54)
8 fq AAq = As ,
(6.55)
where condition (6.42) still applies, with T of Eq. (6.2) being replaced
by T of Eq. (6.45).
In this asymptotic limit the analysis becomes simplified, from a
many-amplitude case to that of only two amplitudes defined by
Eqs. (6.44). The reduced equations (6.43)(6.55) also provide a better understanding of the plateau behaviour. The minimum values of
q which result from these equations are thereby the same as those
obtained from the more elaborate general analysis of 6.6.1.
The plateau analysis also suits an accurate numerical iteration process. Starting from the given values of the constants f0 137.036 and
M 0.00116 defined by Eqs. (5.46) and (5.47), the following set of
equations is considered. First, the ratio
Am
h0 =
As
(6.56)
kq =
(
q )2
.
f 0 h0
(6.57)
kM =
(1 + M )
.
h0
(6.58)
1 +
2)
0
2(
As
f =
=
.
(6.59)
0
8 AAq
Fifth, the flux factor obtained from the last member of this equation
61
value qm
of the normalized charge for a given number of amplitudes of the polar function (6.2).
6.6.3
The simplest case of only two amplitudes a1 and a2 has already been
treated in 6.6.1, thereby resulting in the value q ' 1.01. We now
proceed to a larger number of nonzero amplitudes as follows:
A detailed plateau analysis is first performed on the four-amplitude
case, with its special sub-case of three amplitudes represented by
the line a4 = 0 in Fig. 6.4. The latter clearly demonstrates the
62
Bo Lehnert
plateau behaviour and supports the theory of 6.6.2. With the definition (6.43) and introducing k3 = sin p and k4 = cos p , the
angle p will describe the positions of a3 and a4 at long distances
from the origin a3 = a4 = 0, i.e. along the perimeter of the plateau
region in Fig. 6.4. The self-consistent minimum values of q have
then been found to increase almost linearly from q ' 0.969 for
f ' 1.81 in the three-amplitude case where p = 2 , to q ' 1.03
for f ' 1.69 when p = 4 . Consequently, the plateau of Fig. 6.4
becomes slightly warped, being locally partly below and partly
above the level q = 1 for the self-consistent solutions obeying condition (6.42);
For an increasing number of amplitudes, i.e. more than four, there
is a similar plateau behaviour, and the flux function is also found
to have a form being similar to that demonstrated by Fig. 6.5.
Sweeping again the amplitude parameters at a fixed value f =
= fq = 1.82 of the flux factor, there was hardly any detectable
change in the plateau level87. As an example of a corresponding
qm
' 0.969, being obtained at the plateau in the three-amplitude
case.
The flat plateau behaviour of q at an increasing number of included
amplitudes of the expansion (6.45) can be understood as follows. The
A
last member of Eq. (6.53) includes the ratio A q . Inspection of Eqs. (6.27)
and (6.32) shows that this ratio can be taken to stand for the ratio
between the charge q0 and the magnetic flux 0 . The latter is generated
by the circulating current density (3.3) which is directly proportional to
the charge density. Therefore the considered ratio does not depend on
the total charge, but solely on the profile shape of the charge distribution
in space. Consequently q is expected to have a plateau-like character,
i.e. to become a rather slow function of the higher multipole terms
in the expansion of Eq. (6.45).
6.6.4
63
The reason for the deviation of the deduced minimum charge from the
experimentally determined value is not clear at this stage, but a quantum mechanical correction may offer one possibility for its removal.
Due to the extremely small dimensions of the present configuration, the
quantum conditions are also likely to have an influence in more than
one respect. Here two speculative and preliminary proposals are made,
possibly also to be combined with each other.
A successful quantum mechanical modification of the magnetic moment has earlier been made by Schwinger and Feynman82, as given by
Eq. (5.47). Without getting into the details of the advanced and workconsuming analysis which leads to this correction, it may here be proposed that also the magnetic flux and its related quantum condition
(6.55) would then have to be modified into the form85
As
c
fq =
(1 + ) ,
= ,
(6.60)
f0
8 AAq
where c is a so far unspecified constant. For this proposal to lead
to an agreement with the experimental value q = 1, a modified selfconsistency relation
fq = ff
(6.61)
would have to be satisfied by c ' (4)1/2 . This value is obtained from
the three-amplitude case, within the limits of accuracy of the present
numerical analysis.
The second proposal concerns the quantum mechanical fluctuations.
It can thus be imagined that, at the scale of the present configuration,
the amplitudes (a1 , a2 , a3 , . . .) would become subject to such fluctua
could deviate from its
tions in time. As a result, the minimum value qm
quantum mechanical expectation value <qm >. Since it has been found
in 6.6.3 that q attains values at the plateau which are both below and
64
6.7
Bo Lehnert
(6.62)
65
the circular path of the electron model9. Such a path is instead imposed
by the forced circulation around the axis of symmetry. We therefore
reverse the present effect due to General Relativity by considering
mg to be the equivalent mass which is the cause of the circular motion.
The equivalent gravitation force is then estimated as follows.
Combining the radius (6.63) with the mass density g = wc2s of the
electromagnetic field, the equivalent local gravitation force per unit volume in the radial direction becomes
Gg mg g
g c2
fg '
.
(6.64)
=
2
rg
3rg
Apart from the factor 13 this force can as well be interpreted as the
result of a centrifugal acceleration which generates a force in the outward
radial direction. This is the equivalent extra force to be included into
the set of modified field equations.
The force (6.64) is now compared to the radial electrostatic force
fE = Er
(6.65)
0 E r
3cq q0
= 0 (div E) '
.
4rg3
rg
(6.67)
9c2q q02
16 2 0 rg5
(6.68)
This yields
fE '
and a local force ratio
fg
40 cg c2
m0 rg .
'
fE
9c2q q02
In a first approximation the ratio
cg
c2q
(6.69)
taken to be of order unity. With rg approximately equal to the characteristic radius r0 of the configuration, the average force ratio (6.69)
then reduces for q0 = e to
g
40 c2
<fg >
'
m0 r0 ' 1044 m0 r0
<fE >
9e2
(6.70)
66
6.7.2
Bo Lehnert
A crude first-order estimate of the effect of the equivalent centrifugal force fg of Eq. (6.64) is now made in replacing the steady-state
momentum equation (3.16) by the form
fB ' fE fg + |div 2 S|
(6.71)
with
fB = |
C B| ,
fE = |
E| .
(6.72)
Here fE and fg are directed out of the body of the electron, whereas
fB plays the role of an inward directed confining magnetic force, as
later being explained in detail in 6.8. This implies that part of the
magnetic force now has to balance the force (6.64), and the rest of the
same magnetic force should then be in equilibrium with the additional
forces in Eq. (6.71). We further assume the terms of the steady-state
equation (3.16) to be of the same order of magnitude, as obtained from
a simple dimensional analysis. The final result of this crude estimate is
then that the effect of the force (6.64) can be represented by a reduced
magnetic force (1 g )fB in the unmodified analysis of the previous
6.16.6. A corresponding modification of the field equations is then
obtained in replacing the vector potential A = (0, 0, A) introduced at
the beginning of Chapter 5 by the modified potential
A = (0, 0, A ) ,
A = (1 g )A A + A .
(6.73)
(6.74)
= CA = (sin2 ) DG
G
G0
(6.75)
Iq = f ,
IM = s f + f ,
+ fg , Is = s f (
+ fg ,
Im = f g + h
g + h)
67
(6.77)
(6.78)
where
f = sD(1 + s2 D)(s2 DG) ,
(6.79)
(6.80)
= 2s2 DG ,
h
(6.81)
In a first estimation of the order of magnitude of the present modification we now refer to an earlier simplified analysis which has been based
on the form
(sin )
T = 1 + N
(6.82)
9
68
6.8
Bo Lehnert
The fundamental description of a charged particle in conventional theory is deficient in several respects. Thus, an equilibrium cannot be
maintained by the classical electrostatic forces, but has been assumed
to require forces of a nonelectromagnetic character to be present2325.
In other words, the electron would otherwise explode under the action
of its self-charge.
Here it will be shown that a steady electromagnetic equilibrium can
under certain conditions be established by the present extended theory.
The underlying physical mechanism will first be demonstrated by some
simple examples, and then be followed by its application to the present
axisymmetric electron model in the plateau region.
6.8.1
First consider two parallel line charges in empty space, both with the
charge density per unit length and being separated at the distance r.
They are then subject to the electrostatic repulsion force
fE =
2
20 r
(6.83)
per unit length. But in the case of the present theory, these line charges
also become parallel line currents of the strength
J = c
(6.84)
69
per unit volume in the outward direction. At the same time there is
a current density according to Eq. (3.3), with C in the axial direction.
This generates a magnetic field in the direction, and a corresponding
magnetostatic force
0 c2 2 r
fB =
= fE .
(6.88)
2
Also in this example there is thus a steady electromagnetic equilibrium in which the outward directed electrostatic expansion force is
balanced by an inward directed magnetostatic pinch force.
When proceeding from the straight geometry to a curvilinear configuration, such as that in spherical and axial symmetry, a local balance
cannot generally be realized, but only under certain conditions. This
is due to the fact that, apart from straight cylindrical geometry, the
equipotential surfaces of the electrostatic potential do not generally
coincide with the magnetic field lines determined by the magnetic vector
potential A.
6.8.2
ZZ
G 1 2
Fr = 20 G20
DG +D(s2 DG)
s DG 2 s d d , (6.90)
where s = sin . For the point-charge-like model of Eqs. (6.1)(6.2) with
G = RT , 2 and R 12 at small , we then have
2 DG = D T 2T
(6.91)
(6.92)
70
Bo Lehnert
I+
I
I+
I
= 1.27
at q = 0.98 to
= 0.37
71
Chapter 7
The electrically neutral steady states described in 5.2 will in this chapter be used as the basis for a model of the neutrino. In principle these
states include all the three alternatives Aa, Ab, and Bb of Table 5.1,
but since there is little difference in the outcome of cases Aa and Ab,
the analysis will be limited to Aa and Bb. The model to be developed
here will at least be seen to reproduce some of the basic features of
the neutrino. Since the analysis is restricted to the steady state of a
particle-shaped configuration, thus being at rest, it includes the concept of a nonzero rest mass. The existence of such a mass appears to be
supported by the observed neutrino oscillations. At this stage it can on
the other hand not be expected that the present theory should be able
to describe the occurrence of more detailed and subtle properties such
as the various types of propagating neutrinos, neutrino oscillations, and
the question of the missing right-hand neutrino. Such questions have to
be postponed to a more advanced treatment based on a fully developed
quantum field theory.
The basic theory of electrically neutral particle-shaped states has
already been elaborated in Chapter 5, including general expressions
for the integrated field quantities and the symmetry properties. This
implies that we can now turn to the details of two possible models
derived from the corresponding choices of the generating function89.
7.1
In the earlier performed analysis9 one model has been elaborated which
arises from a convergent radial part and a top-bottom symmetrical
polar part of the generating function of case Aa in Table 5.1, as
given by
G=RT ,
R = e ,
T = (sin ) ,
(7.1)
where 1. We insert the function (7.1) into the forms (5.22)
(5.30), and combine the obtained result with Eqs. (5.18)(5.20), the
relations
Z
1 3 5 . . . (2n 1)
,
C2n =
(sin )2n d =
(7.2)
2 4 6 . . . 2n
0
C2n+1 =
(sin )2n+1 d = 2
73
2 4 6 . . . 2n
,
3 5 7 . . . (2n + 1)
C2n+2
(2n + 1)
=
,
C2n
(2n + 2)
and the Euler integral
Z
2 2
J0 =
e
d =
0
1
22+1
(7.3)
(7.4)
(2 1) .
(7.5)
Jm
15
=
.
Js
38
(7.6)
15h
' 7 1044 [kg m].
152c
(7.7)
We now turn to case Ab in Table 5.1 where the generating function has
a divergent radial part, and a polar part with top-bottom antisymmetry.
A radial part is adopted here which has the same form (6.1) as that of
the electron model in 6.1. Then the earlier discussion on the included
factors also applies here. The forthcoming results in this chapter will be
seen not to depend explicitly on the radial parameter . When r = r0
increases monotonically from r = 0, the radial part of Eq. (6.1) decreases
from a high level, down to R = e1 at r = r0 , and then further to very
small values when r r0 . Thus r = r0 can here be taken as an effective
radius of the configuration.
In an electrically neutral state the expansion (6.2) of the polar part
has now to be replaced by one of general top-bottom antisymmetry with
respect to the equatorial plane. As already shown in 5.2, this leads to a
state having vanishing charge and magnetic moment, but nonzero mass
and angular momentum.
74
Bo Lehnert
7.2.1
For a divergent radial part (6.1) the contributions to the integrals (5.20)
at small lower limits k originate from the ranges close to these limits,
as already pointed out in 6.2. Also here the analysis is analogous
to that of the electron model. This implies that Im and Is are still
obtained from Eq. (6.7), that Eqs. (6.10), (6.11) and (6.14) will apply,
and that Eq. (6.15) refers to k = m, s. Insertion of the obtained forms
into Eqs. (5.18) and (5.19) then yields
0
1
1
r0 G20
Jm ,
c2
2 1 21
m
m0 =
s0 =
0 C
1
1
r0 G20
Js .
c2
2( 1) 2(1)
s
(7.8)
(7.9)
To obtain nonzero and finite values of the mass m0 and the angular
momentum s0 at decreasing radial limits m and s we now introduce
a shrinking effective radius r and a shrinking amplitude factor G0 as
defined by
r = r0 = cr ,
G0 = cG ,
(7.10)
where cr , cG , and are all positive constants and 0 < 1 with as
a smallness parameter. Also here a situation is excluded where , m ,
and s all become exactly equal to zero in the unphysical case of a point
mass with no internal structure. With the introduction of Eqs. (7.10),
expressions (7.8) and (7.9) take the forms
m0 =
s0 =
0
1
1+2
2
c
c
J
,
r
m
G
c2
2 1
21
m
0 C 2 2
1
2(1+)
cr cG
Js 2(1) .
2
c
2( 1)
s
(7.11)
(7.12)
s = (1+)/(1) .
(7.13)
(7.14)
75
h 1 Jm
.
2c 2 1 Js
(7.15)
s = 5/4 ,
(7.16)
i.e. where m and s decrease nearly linearly with . Then the simple
top-bottom antisymmetric form T = cos gives rise to the integrands
Is
Im = 6 cos2 21 cos4 + 18 cos6 (sin ) =
,
sin
(7.17)
which are integrated to Jm ' 0.743, Js ' 0.441, and JJm ' 1.686. We
s
observe that Im then becomes negative within parts of the interval
0 6 6 .
Consequently, relation (7.15) can in a first approximation be written as
m0 r ' 2 1043 [kg m].
(7.18)
7.3
76
Bo Lehnert
We first turn to the result (7.7). The ratio rr between the effective
N
neutrino radius and the nucleon radius is estimated for the given upper
bounds of the neutrino rest mass. This results in ratios of about 106 ,
40, and 0.4 for the electron-neutrino, the muon-neutrino, and the tauonneutrino.
Concerning the large estimated ratio for the electron-neutrino, the
corresponding interaction is then expected to take place between the
short-range nucleon field as a whole on one hand, and a very small part
of the neutrino field on the other. If this small influence on the neutrino
field by a nucleon impact could heal itself by a restoring effect or
by quantum mechanical tunneling, then the effective interaction would
become very weak. This would also apply to impacts with the electrons
being present in the matter structure, since these have an extremely
small radius. If such an imagined interaction becomes relevant, the
neutrino would represent the fog and the matter particles the bullet.
The values of rr for the muon-neutrino and the tauon-neutrino are
N
on the other hand of order unity, and the neutrino mean free paths would
thus become comparatively short when being based on relation (7.7).
7.3.2
According to relation (7.18) the upper bounds of the rest mass will
instead correspond to ratios rr of about 4 106 , 100 , and for the
N
electron-neutrino, muon-neutrino and tauon-neutrino. These ratios can
then become small for all neutrino types when becomes very small.
In such a case the neutrino would instead play the role of the bullet
and the nucleon that of the fog.
77
Even if there does not arise any net electric charge in the present neutrino model, there is still a nonzero local volume force (E+CB). To
attain the state of an integrated force balance, one can proceed in a way
being analogous to that described in 6.8.2 for the electron model. The
polar part T () of the generating function would then have to include an
additional free parameter, to be chosen such as to make the integrated
radial volume force vanish. A corresponding detailed analysis will not
be present here.
7.5
Chapter 8
PLANE WAVES
Because of their relative simplicity, plane waves provide a convenient
first demonstration of the wave types defined in 4.2. In this chapter
their features will be studied, without imposing quantum conditions. An
application is further given on total reflection at the vacuum interface
of a dissipative medium.
8.1
= t + k r ,
(8.1)
(8.2)
B = k E .
(8.3)
With k chosen in the z direction and C located in the plane perpendicular to B, these types are illustrated as follows:
When k E = 0 and k E 6= 0 there is a conventional electromagnetic (EM) wave with a magnetic field according to Eq. (8.3).
The dispersion relation becomes
= kc
(8.4)
79
Fk = c2 kk C ,
(8.6)
C
vp = = k
(8.8)
k
k
and the group velocity is given by
vg =
= C.
k
(8.9)
Thus these velocities both differ from each other and from those
of the EM and S waves. The field vectors E and C have components that are both perpendicular and parallel to the wave normal.
From Eq. (8.3) we further have k B = 0 and E B = 0. Scalar
multiplication of Fk in Eq. (8.6) by C yields C Fk = 0 when
using relation (8.7). Combination of this result with the scalar
product of Eq. (8.6) with C yields E C = 0. Finally, the scalar
product of Eq. (8.6) with E results in E2 = c2 B2 when combined
with Eq. (8.3).
For the EMS wave it is seen that k and E are localized to a plane
being perpendicular to B, and that E and C form a right angle. We
can then introduce the general relation
k E = kE (cos ) ,
(8.10)
where E = |E|. Here the angle stands for the extra degree of freedom
introduced by the nonzero electric field divergence. There is a set of
wave solutions which range for a decreasing from the EM mode given
by = 2 , via the EMS mode in the range 2 > > 0, to the S mode
where = 0. The choice of wave type and of the parameters C and
depends on the boundary conditions and the geometry of the particular
configuration to be considered.
80
Bo Lehnert
81
nontransverse and longitudinal waves. In particular, when a dense medium is dissipative, this leads to questions not being fully understood in
terms of classical electromagnetic theory and Fresnels laws, as pointed
out by H
utt29. The incident and reflected waves in the dense medium
then become inhomogeneous (damped) in their directions of propagation. As a consequence, matching at the interface to a conventional
undamped electromagnetic wave in the vacuum becomes impossible.
This problem is now considered in a frame (x, y, z) where x = 0
defines the vacuum interface. The orientation of the xy plane is chosen
such as to coincide with the plane of wave propagation, and all field
quantities are then independent of z. In the dense medium (region I)
where x < 0 the refractive index is nI n > 1, and an incident (i)
damped EM wave is assumed to give rise to a reflected (r) damped
EM wave. The vacuum (region II) corresponds to x > 0 and has the
refractive index nII = 1. The normal direction of the boundary forms
the angle with the wave normals of the incident and reflected waves.
The wave numbers24 and phases (8.1) of these weakly damped waves
then result in
i,r =
ct n(cos )x + n(sin ) y +
c
(8.11)
+ i
n (cos ) x + (sin ) y
c
with the upper and lower signs corresponding to (i) and (r). Here the
1
damping factor is denoted by = 2
1, with standing for the
electric permittivity and for the electric resistivity in region I. For the
phase of a transmitted wave in region II the notation
(qt x + st y)
(8.12)
t = ( ct + pt x + rt y) + i
c
c
is introduced where all quantities (pt , rt , qt , st ) are real.
The possibility of matching a transmitted EM wave to the incident
and reflected waves is first investigated. This requires the phases (8.11)
to be matched to the phase (8.12) of the transmitted wave at every
point of the interface. This condition can be written as
rt = n > 0 ,
st = n > 0 ,
n = n(sin ) ,
(8.13)
(8.14)
(8.15)
82
Bo Lehnert
The transmitted wave should further travel in the positive x direction, into region II, and this also applies in the limit where the angle
of its normal with the vacuum interface approaches zero during total
reflection. Thus, pt > 0. Eqs. (8.15) and (8.13) then combine to the
condition
n2
qt =
< 0.
(8.16)
pt
For total reflection, however, there should be no flow of energy into
region II, and the transmitted wave must represent a flow directed parallel to the interface, thereby being limited in amplitude to a narrow
layer at the vacuum side of the interface24. This excludes the negative
value of qt given by Eq. (8.16), as well as the form (8.12) for an EM
wave. Therefore it does not become possible to match the inhomogeneous (damped) EM waves in region I by a homogeneous (undamped)
EM wave in region II. This agrees with an earlier result by H
utt29.
Turning instead to the possibility of matching the incident and reflected EM waves to transmitted EMS waves in the vacuum region, we
consider the two cases of parallel and perpendicular polarization of the
electric field of the incident wave. The velocity C of an EMS wave is
now given by
(8.19)
According to Eqs. (8.19) and (8.13) matching of the phase by a transmitted EMS wave becomes possible when
qt = st tan = n tan .
(8.20)
In the case of total reflection the velocity vector C of Eq. (8.17), and
the corresponding current density (3.3), can then be directed almost
parallel with the interface x = 0, i.e. when | cos | 1, | sin | ' 1 and
tan < 0. The EMS wave in region II is then matched to the amplitudes
of the EM waves in region I which slightly decrease in the y direction.
This also implies that qt in Eq. (8.20) can be made positive and large for
weakly damped EM waves in region I, i.e. when the resistivity is small
and the damping factor large. Even a moderately large qt > 0 provides
the possibility of a transmitted energy flow along the interface, within
83
a narrow boundary layer and for an EMS wave amplitude that drops
steeply with an increasing distance from the interface.
As a next step the electric and magnetic fields have to be matched at
the interface. This raises three questions being in common with those
of conventional theory24. The first issue is due to the expectation that
the transmitted and reflected waves are no longer in phase at the surface x = 0 with the incident wave. The second question concerns the
amplitude ratio of the reflected and incident waves which must have a
modulus equal to unity. This is due to the expectation that no energy
can be lost during the instantaneous reflection process. This question
also leads to the third issue that concerns the energy flow of the transmitted wave. As already being pointed out, this flow has to be directed
along the interface and be localized to a narrow boundary layer.
To meet these requirements we observe that the wave number and
the phase are coupled through the angles of the velocity form (8.17).
In this way the angle of any transmitted EMS wave in region II can be
expressed in terms of the angles and .
The details of the deductions to follow are given elsewhere9,17, and
are summarized as follows:
For inhomogeneous (damped) incident EM waves the necessary
matching of the phases at the interface can be provided by nontransverse EMS waves in the vacuum region;
The transmitted EMS waves are confined to a narrow layer at the
vacuum side of the interface, and no energy is extracted from the
reflection process;
A far from simple question concerns the magnitude of the damping factor which in physical reality forms a limit between the
analysis of damped and undamped incident waves. In most experimental situations the ratio 1 between the damping length and
the wave length is very large, and this makes it difficult to decide
when the present analysis on inhomogeneous incident waves becomes relevant. It should first become applicable at large damping
factors, but this requires large initial amplitudes to give rise to a
detectable reflected wave.
Chapter 9
During several decades a number of investigators have discussed the nature of light and photon physics, not only in relation to the propagation
of plane wave fronts but also to axisymmetric wave packets, the concept
of a photon rest mass, the existence of a magnetic field component in
the direction of propagation, and to an associated angular momentum
(spin) of the photon in the capacity of a boson particle.
The analysis of plane waves is relatively simple. As soon as one
begins to consider waves varying in more than one spatial dimension,
however, new features arise which complicate the analysis. This also applies to the superposition of elementary modes to form wave packets. In
this chapter dissipation-free axially symmetric modes are investigated
on the basis of the present theory. A corresponding wave packet configuration then results in a photon model, being one of the possible
representations of an individual photon concept.
In analogy with the previous analysis on axisymmetric equilibria,
we will also here search for a model where the entire vacuum space
is treated as one single entity, without introducing internal boundaries
and boundary conditions.
The aim is here to elaborate a model of the individual photon in the
capacity of a boson particle with an angular momentum(spin), propagating with preserved geometry and limited extensions in a defined
direction of space. This leads to the concept of cylindrical waves and
wave packets.
The first parts of this chapter will deal with axisymmetric wave
modes, whereas screw-shaped modes are treated to some extent in its
later parts.
9.1
A cylindrical frame (r, , z) is introduced where is an ignorable coordinate. In this frame the velocity vector is assumed to have the form
C = c (0, cos , sin )
(9.1)
85
with either sign of cos and sin , and a constant angle . We further
define the operators
2
2
1
1 2
+
+
,
r2
r r z 2
c2 t2
+ c (sin )
.
D2 =
t
z
D1 =
(9.2)
(9.3)
The basic equations (4.2) and (4.3) are then represented by the system
D1 2 Er =
(div E) ,
(9.4)
r
r
1
D1 2 E = (cos )
(div E) ,
(9.5)
r
c
t
D1 E z =
+ (sin )
(div E) ,
(9.6)
z
c
t
and
D2 (div E) = 0 .
(9.7)
Thereby Eq. (9.7) originates from Eq. (4.3) and does not introduce
more information than that already contained in Eqs. (9.4)(9.6). By
further defining the operator
2
1 2
D3 = 2 2 2
(9.8)
z
c t
Eq. (9.4) reduces to
2 Ez
D3 Er =
.
(9.9)
r z
Combination with Eq. (9.5) then yields
1
1
2
D3 D1 2 E = (cos )
D1 Ez .
(9.10)
r
c
z t
86
Bo Lehnert
In the case of a vanishing electric field divergence, Eqs. (9.4)(9.6) reduce to those of an axisymmetric EM mode. The separation constant
of the z- and t-dependence is
2
2 = k 2
.
(9.11)
c
Then the phase and group velocities vp and vg become
r
2
vp = = c 1 2 ,
(9.12)
k
k
vg =
c
= q
k
1
2
k2
(9.13)
Here 2 > 0 results in vg2 < c2 with solutions including Bessel functions of the first and second kinds24.
As a next step we assume 2 to vanish which corresponds to a phase
and group velocity equal to c. Then Eqs. (9.4)(9.6) reduce to
Ez = 0 ,
D 2 (Er , E ) = 0 ,
D
(9.14)
where
2
= + 1 .
D
(9.15)
2
= k1 + k2 ,
E
where kr1 , kr2 , k1 , k2 , kz1 and kz2 are constants. The magnetic field
is obtained from Eq. (3.7).
A similar divergence of the field at the axis = 0 and at large
was already realized by Thomson26 and further discussed by Heitler23,
as well as by Hunter and Wadlinger27. It leaves the problem with the
radial dependence unresolved. Thus the field does not converge within
the entire vacuum space, and cannot be made to vanish at large radial
distances as long as kr1 , k1 and kz1 differ from zero. All parts of the
87
(9.17)
2
+ 2 E = 0.
(9.18)
x2
y
With separable solutions of the form f = X(x) Y (y) this yields
X = cx1 ekx x + cx2 ekx x ,
ky y
and
Y = cy1 e
+ cy2 e
ky y
kx2 + ky2 = 0 ,
(9.19)
(9.20)
(9.21)
88
Bo Lehnert
9.1.2
1
2
2
+
k (cos ) E =
r2
r r r2
(9.23)
2
1
2
2
= (tan )
+
k (cos ) Ez .
r2
r r
We can now introduce a generating function
G0 G = Ez + (cot ) E ,
G = R () ei(t+kz) ,
(9.24)
1
1
(1 2 D) G =
+ iAr ,
(cos )2
r0
(9.26)
E = G0 (tan )2 DG = iA ,
(9.27)
Ez = G0 (1 2 D) G = ik + iAz
(9.28)
and
1
(9.29)
2 DG = ikA ,
c (cos )
1
1 Az
B = iG0 (sin )
, (9.30)
(12 D) G = ikAr
2
c (cos )
r0
1
1 2
1
Bz = iG0
(9.31)
+
( DG) =
(A ) .
c (cos )
r0
Br = G0
89
Here the parameter should not be confused with the polar angle in
the spherical frame of coordinates applied in Chapters 5 and 6. These
expressions are reconfirmed by insertion into the basic equations. The
obtained result demonstrates that the electromagnetic field of the normal mode is determined by one single generating function (9.24), in
analogy with the function (5.6) of the steady state, and that the three
equations (9.4)(9.6) are not strictly independent of each other. This
is a special feature of the present theory. It makes possible a choice
of generating functions and corresponding modes which are physically
relevant within the entire vacuum space, and which become consistent
with the imposed quantum conditions.
Due to the dispersion relation (9.22) the EMS mode propagates at
phase and group velocities which are smaller than c. Not to get into
conflict with the experimental observations by Michelson and Moreley,
we limit ourselves for a positive cos to the condition
0 < cos 1 ,
(9.32)
90
Bo Lehnert
v2
= cos ,
c2
v = 0, 0, c (sin ) .
(9.33)
=
,
k 0 = k ,
0 = k 0 r0 = .
0
z
z
r0 = r ,
z0 =
(9.34)
(9.35)
G = R () ei(t+kz) = R () eik z G0 ,
(9.36)
DG = D0 G0
(9.37)
1
1
1
0
+ v B,
E = E
1 (
z E) z
(9.38)
1
1
1
E
0
v 2 ,
B = B
(9.39)
1 (
z B) z
c
= (0, 0, 1). This results in the components
where z
Er0 = iG0 (0 )1
(1 2 D0 )G0 ,
(9.40)
E0 = 0 ,
(9.41)
Ez0 = G0 (1 2 D0 ) G0 ,
(9.42)
and
Br0 = G0 c1 2 D0 G0 ,
B0 = 0 ,
Bz0 = i G0 (c 0 )1
(9.43)
(9.44)
(2 D0 G0 ) .
(9.45)
91
(0 = 0)
(9.46)
in K . This supports the adopted form (9.1). In the rest frame K 0 the
current density (2.8) has a component in the direction only, and it
circulates around the axis of symmetry, thereby being connected with
electric and magnetic fields which are purely poloidal, i.e. located in
planes running through the same axis. This situation is similar to that
of the bound steady equilibrium state of Chapter 5, with its static
electric and magnetic fields.
9.2
(9.47)
The normal modes are superimposed to result in a wave packet having the amplitude
2
2
k
Ak =
ez0 (kk0 )
(9.48)
k02
in the interval dk and being centered around the wave number k0 . Further 2z0 represents the effective axial length of the packet. In the integration of the field components (9.26)(9.31) the expression
Z +
P =
k Ak eik(zvt) dk ,
v = c (sin )
(9.49)
i z
,
2z0
z = z vt
(9.50)
z
k0
1
z 2
P =
exp
+ ik0 z 1 + f
,
, (9.51)
k0 z0
2z0
z0 k0 z0
92
Bo Lehnert
0
= k 1z = 2z
1, where 0 = 2
k0 is the average wave length of the
0 0
0
packet. Under this condition the contribution from f to the form (9.51)
can be neglected with good approximation, as also being seen from a
detailed evaluation of the integral (9.49). On the other hand, the wave
packet model would also apply when the axial extension 2z0 is of the
order of one wave length 0 , but this leads to an extremely broad line
width not being reconcilable with experiments.
With the notation
g0
z 2
E0 E0 (
z) =
exp
+ ik0 z ,
(9.52)
k0 r0
k0 z0
2z0
k
k0
r = iE0 R5 + (00 )2 R2 ,
E
z = E0 0 (cos )2 R4 + (00 )2 R1 ,
E
r =
B
1
,
E
c (sin )
= 1 (sin ) E
r ,
B
c
z = i 1 E
0 (cos ) R8 (00 )2 R7 ,
B
c
(9.53)
(9.54)
(9.55)
(9.56)
(9.57)
(9.58)
d
1
d
1
R7 =
+
R1 ,
R8 =
+
R3
d
d
R1 = 2 R ,
R2 =
(9.59)
(9.60)
(9.61)
93
(9.62)
when the real parts of the forms (9.24), (9.49), and (9.52) are adopted.
, E
z , Br ) are symmetric and (E
r , B
, B
z ) are antisymThen G and (E
metric functions of z with respect to the centre z = 0.
9.3
Z Z +
+
= 20
r
Er d
z + r Ez dr = 0 ,
r
(9.63)
r is antisymmetric and E
z symmetric with respect to z.
because E
The integrated magnetic moment is related to the component
c (cos )
j = C = 0 (div E)
(9.64)
Z +
Z
(9.65)
2 +
= 0 c (cos )
r
r
Er d
z + r Ez dr = 0
r
94
Bo Lehnert
for the same symmetry reasons as those applying to the charge (9.63).
It should be observed that, even if the net magnetic moment vanishes,
the local magnetic field remains nonzero.
As in the case of the neutrino model of Chapter 7, the present photon model can have a structure with both positive and negative local
contributions to the charge and the magnetic moment, but these end
up to zero when being integrated over the total volume.
9.3.2
Total mass
g0
5/2
2 z0
2 a0 g02 .
a0 = 0
ck02 z0
(9.69)
(9.70)
For this wave packet the energy relations by Planck and Einstein
also combine to
hc
2
mc2 = h0 '
,
0 =
,
(9.71)
0
k0
where use has been made of the dispersion relation (9.22) applied to the
wave packet as a whole, i.e.
c
(9.72)
0 = 20 ' k0 c ,
0 =
.
0
95
Angular momentum
f = (E + C B) ,
(9.74)
1
z .
r ) + E
=
(rE
0
r r
z
(9.75)
S
B
s=r 2 ,
S=E
(9.76)
c
0
apply for the density s of the angular momentum, where r is the radius
vector from the origin and S is the Poynting vector. The density of
angular momentum now becomes
z B
r E
r B
z ) ' 0 r E
r B
z ,
sz = 0 r (E
(9.77)
96
Bo Lehnert
when condition (9.32) applies and use is made of Eqs. (9.53), (9.55),
(9.56), and (9.58). The total angular momentum is given by
Z
Z + Z
r B
z dr d
s = sz dV = 20
r2 E
z,
(9.78)
h
,
2
Z
Ws =
2 R5 R8 d
(9.79)
9.3.4
97
R () = e
(9.80)
R () = e1/ ,
(9.81)
Ws
.
(9.83)
(9.84)
0
.
(cos )
(9.85)
98
Bo Lehnert
99
(9.79) to shrink to very small but nonzero values, as the lower limits
m and s approach zero. This is attained by introducing the relations
r0 = cr ,
g0 = cg ,
cr > 0 ,
cg > 0 ,
(9.89)
(9.90)
m = a0 5 c2g
(9.91)
(9.92)
where the average wave length 0 of the packet has been defined in 9.2.
To obtain finite values of both m and s, it is then necessary to satisfy
the conditions
m = / ,
s = (2+1)/(21) .
(9.93)
We are here free to choose = 1 by which
s ' m =
(9.94)
with good approximation. The lower limits of the integrals (9.87) and
(9.88) then decrease linearly with and the radius r0 . This forms a
similar set of geometrical configurations which thus have a shape being independent of m , s , and in the range of small .
The ratio of expressions (9.91) and (9.92) finally yields an effective
photon diameter
0
2 r =
.
(9.95)
(cos )
It should be observed that the wave packet diameters of Eqs. (9.95)
and (9.85) both become independent of the particular values of the parameters and . As compared to the relatively large photon diameter
(9.85) obtained for a convergent generating function, the diameter (9.95)
based on a divergent such function can shrink to very small dimensions.
This is the case even when cos 1 provided that cos . Then the
photon model becomes strongly needle-shaped in its transverse directions. This has no counterpart in the conventional axisymmetric solutions of 9.1.1, which in addition would consist of divergent integrated
field quantities due to the contributions from large radial distances.
100
9.3.6
Bo Lehnert
The wave packets studied in 9.3.4 and 9.3.5 are deduced from radial
parts (9.80) and (9.86), having large values of and where
cos 1. The dominating field components of the normal modes
(9.26)(9.31) and (9.40)(9.45) in the laboratory and rest frames, K
and K 0 , then become related by
Er = cB =
Er0
cBz0
=
.
cos
cos
(9.96)
A corresponding field energy density (3.25) is obtained for the superimposed normal modes which form the spectrum of a wave packet. For
the latter there is a Lorentz contraction in the direction of propagation
in the moving frame K, as defined by Eqs. (9.33) and (9.34). The total
energies of the wave packets in the frames K and K 0 are then related by
Z
Z
1
m0 c2
mc2 = wf dV =
wf0 dV 0 =
,
(9.97)
cos
cos
where m0 is the rest mass, in accordance with Eq. (9.73).
This result has earlier been obtained in an alternative way18, for a
convergent radial part, and on the basis of the source energy density
(3.26). Here relations (9.96) and (9.97) can also be recovered in terms
of the average field components (9.53)(9.61) in K and of the corresponding components18 in K 0 .
9.4
The present photon models have features being associated with a number of crucial questions that have been subject to extensive discussions
in the current literature. Some fundamental properties of the photon
models resulting from conventional theory and from the present approach are listed and compared in Fig. 9.1.
9.4.1
The possible existence of a nonzero photon rest mass was first called attention to by Einstein40, Bass and Schrodinger41, and de Broglie and
Vigier31. It includes such fundamental points as its relation to the
Michelson-Morley experiment, and its so far undetermined absolute
value.
The phase and group velocities of the present non-dispersive wave
packets become slightly smaller than the velocity of light, as expressed
by condition (9.32) and Eq. (9.22). Thereby the velocity constant c can
101
Figure 9.1: A comparison between the individual photon models of conventional theory and of the present approach.
v+w
.
1 + vw
c2
(9.98)
Here w = 104 c for the parallel or antiparallel directions of incoming light at the surface, as counted with respect to the orbital motion.
With the condition (9.32) the departure from c of the recorded photon
velocity would then be given by
u
1
2w
1 ' (cos )2 1
.
(9.99)
c
2
c
For the value cos 6 104 corresponding to a photon rest mass
m0 < 0.74 1039 kg, a change in the eight decimal of the recorded velocity would hardly become detectable. For a rest mass in the range
1068 < m0 < 1045 kg considered by Evans and Vigier4, this change
even becomes much smaller. Moreover, when turning from a direction
where wc = + 104 to the opposite direction, the change in 1 uc becomes
as small as about 1012 . Consequently, there should not be observed
any noticeable departure in recorded velocity and deviation from an
102
Bo Lehnert
103
The local electric charge density of the present theory can have either
sign. In the models of 9.3, the total integrated charge still vanishes
according to Eq. (9.63). To obtain zero net charge it does therefore not
become necessary to assert that the photon is its own antiphoton.
The total integrated magnetic moment vanishes as well due to
Eq. (9.65). Nevertheless there is a nonzero local and helical magnetic
field which has a component in the axial direction of propagation. This
component is invariant to Lorentz transformations with respect to the
same direction, and it thus exists as a time-independent component in
the rest frame. These features are reconcilable but not identical with
those of the theory by Evans and Vigier4. The helicity of the photon
magnetic field in the laboratory frame has further been considered by
Evans and by Dvoeglazov in the volume by Evans et al.7
As shown by Eq. (9.79) a nonzero angular momentum only becomes
possible when there is a nonzero rest mass (9.73) with an associated
factor cos . In its turn, this factor is related to a nonzero axial magnetic
field component of the helical field configuration. For a vanishing rest
mass we would thus be back to the conventional case of 9.1.1 where
there is no axial magnetic field.
9.4.3
104
Bo Lehnert
an integrating part of the total energy. Such a subdivision into a particle and an associated pilot wave is therefore not necessary in the present
case where the wave packet behaves as an entity, having both particle
and wave properties at the same time.
A de Broglie wave length = phz related to the axial relativistic momentum pz can also be deduced from the present theory. Thereby there
are contributions to the total energy associated with the two energy
flows along and around the axis.
Turning to the effective photon diameter of these wave packet models, we first notice that a convergent generating function leads to an
individual photon diameter (9.85) which in some cases is rather limited,
but still does not become small enough to match atomic dimensions.
As an example, an average wave length 0 = 3 107 m and a factor
cos = 104 which makes the velocity v deviate from c by 5 109 only,
results in a diameter 2 r ' 103 m. On the other hand such a diameter could become reconcilable with that of a radially polarized dense
multiphoton beam.
A divergent generating function can on the other hand result in a
very small photon diameter (9.95) for sufficiently small , then having
the size of atomic dimensions such as the Bohr radius. This becomes
reconcilable with the ability of the photon to knock out an electron from
an atom in the photoelectric effect. It is consistent with the proposed
needle radiation of energy quanta with a directed momentum as obtained by Einstein95. Also here a proposal on two limits of the photon
diameter can be forwarded, in analogy with 6.7.3 and applying to the
part of the radiation which circulates around the axis of symmetry.
In the two-slit experiment by Tsuchiya et al.22, a light source of low
intensity was used to emit single photons at each instant. The photons
could pass any of the two narrow slits of a small separation distance
at a first screen S1 shown in Fig. 9.2, and then impinge on a second
screen S2 . Each single photon produces one dot-shaped mark at S2 , like
that formed by a needle-shaped bullet. At the same time each single
photon appears to have passed through both slits, and this results in
an interference pattern at S2 . The pattern becomes visible in the form
of a large number of marks which appear after some time when many
photons have reached S2 .
From the quantum mechanical point of view and due to the uncertainty principle, it becomes impossible to determine through which
slit the photon passes, i.e. without destroying the diffraction pattern33.
This is due to the change of momentum which would occur on account
of an applied detector.
105
2t
mark. When many photons have passed in energy. The latter
uncerthe system after each other, the resulttainty then comes out to be
ing dot-shaped marks form an interference
of the order of half the enpattern on S2 . The separation distance between the slits is strongly exaggerated in ergy h0 , thus being carried
on the average by each imagthe figure.
ined photon part.
Recently Afshar et al.96 have reported over a simultaneous determination of the wave and particle aspects of light in a two-slit welcherweg experiment. Their results show that the single photon has particle
and wave properties simultaneously. A particle-like photon thus passes
106
Bo Lehnert
one of the pinholes, but an interference pattern can still be formed. Also
this second interpretation appears to be consistent with the present theoretical model. The sampling can possibly be related to the long axial
length of the wave packet and the ocurrence of a precursor at its front24.
In the two-slit experiments the present needle-shaped wave packet
model has an advantage over a plane wave model. The particle feature
in the form of a very narrow needle-shaped transverse diameter makes
the wave packet model reconcilable with the observed dot-shaped marks.
At the same time the wave nature of the axisymmetric packet makes
it possible for interference phenomena to occur, in the same way as
for a plane wave. Two parts of such a wave packet which are 180
out of phase and meet at a point P on S2 should namely also cancel
each other by interference. This dualism and ability of interference
becomes particularly obvious in the zero line width limit k 1z 0 where
0 0
the spectral distribution of Eq. (9.48) reduces to that of the elementary
normal mode of 9.1.
The geometrical structure of the individual localized photon of the
present theory is not in conflict with the concept of a quantum mechanical wave function. The latter represents the probability distribution of
a photon before its position has been localized through a measurement.
9.4.4
The photon is the field quantum being responsible for the electromagnetic interaction. Likewise the weak field interaction acquires quanta,
in the form of the W+ , W , and Z0 bosons21. This raises the question whether an analogous Proca-type equation applied to the weak
field case could result in axisymmetric solutions being similar to those
deduced here for the electromagnetic field. This may provide the weakfield bosons with a nonzero rest mass, thereby arriving at a possible
alternative to the Higgs particle concept.
9.4.5
The analysis performed so far may debouch into the idea that there exist
different quantum states (modes) of the photon representing different
solutions of the same basic equations. In a way this would become analogous to but not identical with the Copenhagen interpretation. Almost
instantaneous transitions between these states also become imaginable.
A similar situation has already been observed for neutrino oscillations
which are associated with a nonzero neutrino rest mass. Consequently,
it is here proposed that analogous photon oscillations can exist, by
which rapid transitions between the various plane and axisymmetric
107
photon modes take place, under the constraints of total energy and angular momentum conservation. In this way the photon could behave
differently in different physical situations. Such a preliminary proposal
has, however, to be further investigated.
In a first attempt we outline some types of transition which could
make it possible for the photon to interact on the atomic scale, such as
in the photoelectric effect, as well as in other situations which require
its energy to become spatially concentrated:
A transition from the axisymmetric mode of 9.3.4 having a moderately large photon diameter to that of the needle radiation
described in 9.3.5 is at least expected to face no difficulties with
the conservation laws. A well-defined energy threshold is then
preserved in the photoelectric interaction;
A transition from an equivalent plane wave to any of the axisymmetric modes of Sections 9.3.4 and 9.3.5 is more complicated, since
the plane wave has no angular momentum. For such a transition
to become possible, it must occur as a three-mode event, similar to that of pair formation. The energy of the plane wave then
has to be shared by two axisymmetric modes of either or both
types given in the same sections, and the angular momenta of
these modes must be antiparallel. To obtain a well-defined energy
threshold, additional requirements have to be imposed. If there is
only one mode of needle-shape involved, almost all the plane-wave
energy has to be shared by this mode. But if there are instead two
such modes, the energy should either be carried almost entirely
by one of them, or both modes have to take part simultaneously
in the energy transfer of the photoelectric interaction.
The proposed process of photon oscillations can become more specific in an example where a plane wave or a plane wave packet (p ) is
assumed suddenly to decay into two needle-shaped modes, (1 ) and (2 ),
of the type deduced in 9.3.5. Here we limit ourselves to given values
of the parameters cos , of Eq. (9.89), and of the integrals (9.79) and
(9.87). The components of the velocity vector (9.1) are now chosen
to be antiparallel, as given by
C1 = c (cos ) = C2 ,
(9.100)
1
1
hc
+
,
(9.101)
hp =
2 (cos ) r01
r02
108
Bo Lehnert
where use has been made of Eq. (9.95). Since the axial momentum is
equal to h
c according to de Broglie, its balance becomes satisfied by
Eq. (9.101) as well. For the balance of the angular momentum, the
contributions from the two axisymmetric modes have to cancel each
other. With relations (9.100) this leads to
h 01 r01 = h02 r02 ,
(9.102)
as obtained from combination of Eqs. (9.79) and (9.69). From the results
(9.101) and (9.102) it is then seen that almost all the energy of the plane
wave can be transferred to the needle-shaped mode (1 ), say, in the case
of a very large ratio rr02 . This then implies that mode (2 ) will possess
01
a vanishingly small energy which is spread over a larger radius r02 in
space, as compared to that of the mode (1 ).
The idea that the photon can exist in different states is also in a way
related to the different possible forms of the electromagnetic energy
density. Thus the field energy density (3.25) is evidently associated
with the local energy in space of a propagating electromagnetic EM
wave. On the other hand, when considering the axisymmetric state of
an EMS wave packet, we can as well use the source energy density (3.26)
when deducing the total mass.
9.4.6
With a nonzero rest mass one would first expect a photon gas to have
three degrees of freedom, i.e. two transverse and one longitudinal. Then
Plancks radiation law would become altered by a factor of 23 , in contradiction with experience. A detailed analysis based on the Proca
equation shows, however, that the part of the field being associated
with the rest mass, the related angular momentum, and with the axial magnetic field component cannot be involved in a process of light
absorption4,6. This is also made plausible by the axisymmetric wave
packet model in which the axial part of the magnetic field behaves as a
time-independent component in the rest frame, i.e. without having the
features of a propagating wave. It is further observed that transverse
photons cannot penetrate the walls of a cavity, whereas this becomes
possible for longitudinal photons which do not contribute to the thermal
equilibrium41. Consequently, Plancks law is recovered in all practical
cases, also when there is a small nonzero photon rest mass4,6.
The equations of state of the photon gas have further been considered by Meszaros58 and Molnar59 who found that Plancks distribution
and the Wien and Rayleigh-Jeans laws cannot be invariant to adia-
109
Light beams
Density parameters
1/3
1
h c2
=
(cos
)
(9.106)
,
p 40
1/
110
Bo Lehnert
where the upper factor refers to Eq. (9.85) and the lower to Eq. (9.95).
Multiphoton states27 are not considered here, but could somewhat modify the analysis.
Since r is a rather sharply defined radius according to the previous
deductions, there is a critical value c ' 1 of the ratio (9.103) for a
beginning transverse packet overlapping, as given by
c
1
(p 40 )
(cos )
There is in principle also a longitudinal (axial) overlapping of
consecutive wave packets situated on exactly the same axis. Such packets can match their phases and combine into more elongated wave trains
having a more narrow line width. This would have no additional consequences on the beam formation and does not require further consideration.
Transverse packet overlapping implies on the other hand that the
individual photon fields would interfere with each other. The axisymmetric solutions of an individual wave packet then break down, and the
analysis of Sections 9.19.4 in no longer applicable.
9.5.2
(9.108)
111
beam with no overlapping the photon wave packets all carry their own
spin. A high-density beam which has been converted into a plane wave
system would on the other hand have no angular momentum, as long
as the wave has infinite transverse extensions. The angular momentum
of a beam with a finite transverse cross-section can on the other hand
still be preserved. This is due to the transverse derivatives which exist at the boundaries of the beam, and which contribute to an angular
momentum23,97. The angular momenta which would have existed for the
individual photons within the volume of the beam are then imagined
to become substituted by the momentum generated at the bounding
surface. Such a situation somewhat resembles that of gyrating charged
particles in a magnetized and bounded plasma body. There the diamagnetic contributions from all the freely gyrating particles within the
plasma volume cancel each other, and become substituted by a diamagnetism produced by an electric current which circulates along the
plasma boundary.
Thus, a light beam with a spatially limited cross-section and an
angular momentum can in principle be described by the present revised theory19, in analogy with the deductions of 9.3.3 which apply to
the individual photon. Then a rectangular frame becomes appropriate
to the study of a linearly or elliptically polarized beam core. Such a
study applies as well to the limit of an individual photon with a linearly
polarized core and with an internal structure. But it does not deal
with the entangled quantum states of two interacting photons. Also
in this case conventional theory merely leads to a vanishing angular
momentum.
9.5.3
112
9.6
Bo Lehnert
() ei(t+m+kz)
E=E
E()
eim ,
=
.
(9.109)
r0
The form (9.109) represents the general case of cylindrical waves to
be treated here, and like the strictly axisymmetric case of 9.1 it will
lead to nearly radially polarized electric field solutions. Thus elliptically
polarized dense plane-wave beams are not included in this concept. Here
the phase factor m represents screw-shaped(twisted) modes at a given
time t. The parameter m
is a positive or negative integer, the various
values of which represent different states of twisted photons or light
beams. For a single corkscrew-shaped beam with m
= 1, the axial
113
1
2
1 2
D1 2 + 2
Er 2
E =
(div E) ,
(9.110)
2
r
r
r
r
1
1 2
2
E + 2
Er =
D1 2 + 2
r
r 2
r
1
1
=
+ (cos )
(9.111)
(div E) ,
r c
t
1 2
D1 + 2
E
=
+
(sin
)
(div E) ,
(9.112)
z
r 2
z
c
t
and
+ c (cos )
+ c (sin )
(div E) = 0 ,
t
r
z
where
div E =
1
+
r
r
Er +
E +
Ez .
r
z
(9.113)
(9.114)
(9.115)
114
Bo Lehnert
m
1
2
2
E
+
+
(iE ) + k
(iEz ) = 0 ,
+
k
r
r2
c2
r r
r
r
(cos )
+
Er +
r r
r
c
r
r
2
1
1
2
2
+
+
k
+
(cos
)
(iE ) +
r2
r r r2
c2
c
r
hm
i
+k
(cos ) (iEz ) = 0 ,
r
c
h
i
1
m
k (sin )
+
Er + (iE ) +
c
r
r
r
2
h
i
2
1
m
+
2 +
k (sin ) (iEz ) = 0 ,
+
r2
r r
r
c c
and Eq. (9.113) turns into
h
i
m
k c (sin )
c (cos ) (div E) = 0 .
r
9.6.1
(9.110)
(9.117)
(9.118)
(9.119)
(9.120)
In the case of the conventional EM mode where div E and C vanish and
there is no photon rest mass, we have = kc and Eqs. (9.109)(9.113)
reduce to the set
2
1
1
2m
2
+
(1 + m
) (Er , iE ) 2 (iE , Er ) = 0 , (9.121)
2
2
1
m
2
+
2 Ez = 0 .
(9.122)
2
1
m
(Er ) + (iE ) .
(9.123)
(9.124)
Ez = (c1z m
+ c2z m
) eim ,
115
Er = c1r 1m
+ c2r (1m)
eim = iE , (1 m
6= 0) (9.125)
when 1 m
6= 0. Here it is observed that m
= 0 brings us back to
the case already discussed in 9.1.1, as given by Eqs. (9.16) and (9.17).
When on the other hand 1 m
= 0 the solution becomes
Er = [ c1r0 + c2r0 ln] eim = iE ,
(1 m
= 0) .
(9.126)
However, all the general solutions (9.124)(9.126) are either divergent at the origin or at large distances from it. It is further seen from
Eq. (9.115) that the volume forces vanish in the conventional limit. The
axial component of the angular momentum per unit volume is then
given by
(r S)z
sz =
(9.127)
= 0 r (Ez Br Er Bz ) ,
c2
where r is the radius vector from the origin. Thus, a nonzero angular
momentum can only arise for nonzero field components (Er , Ez ) and
(Br , Bz ). But such components always become divergent in space, and
this is not acceptable from the physical point of view. Moreover, and
what again becomes a serious additional constraint as in the case of
Eq. (9.17), the electric field divergence has to vanish for all (, m,
k),
such as for an arbitrary spectrum of a wave packet. This makes the
component Ez disappear, in the same way as Bz . Then the Poynting
vector has only a component in the direction of propagation, and there
is again no spin with respect to the z axis. The same result has also
been obtained by Stratton24 in the limit = kc.
Consequently, a photon model based on the conventional screwshaped mode, and on an arbitrary value of the parameter (9.11), would
also lead to the unacceptable result of infinite field energy. This can
only be avoided in a wave guide where there is a limiting wall.
9.6.2
When cos = 0 and div E 6= 0 the solution of Eq. (9.113) has the form
of a dispersion relation = kc which corresponds to a vanishing rest
mass. Eq. (9.110) then reduces to
k r0
Er =
+ 1 (iE ) i 2 2
Ez .
(9.128)
m
116
Bo Lehnert
Ez
mc
(9.129)
iE = E0 e eim
(9.130)
with positive values of and . Eq. (9.128) then yields
mE
r
= + 1 .
iE
(9.131)
flat maximum the resulting electric field then becomes somewhat like a
linearly polarized wave. The configuration of Eq. (9.130) can further
be limited to a narrow region around the z axis, by choosing a sufficiently
large .
A spectrum of this elementary mode with a vanishing spin can finally
be made to form a wave packet of finite axial length. The corresponding
total energy and mass can then be obtained from the energy density
(3.25), and be quantized in terms of the frequency 0 = c in the case
0
of a narrow line width.
9.6.3
117
2 2
(9.133)
+
k
E
+
+
(iE ) + k
(iEz ) = 0 ,
r
r2
r r
r
r
1
1 2
1
m
k 1
+
Er +
r r
r
2
r
r
2
1
1
m
1 2
2 2
+
+
(iE ) +
r2
r r r2
r
2
1
+k
(9.134)
k 1 2 (iEz ) = 0 ,
r
2
1
m
1
m
2
2
k
+
Er + (iE ) +
+
(iEz ) = 0 . (9.135)
r r
r
r2 r r r2
The lowest order approximation of small is now studied for these
equations. Thereby a negligible axial component Ez becomes reconcilable with a magnitude being of the order of 2 as compared to the
components Er and E . As will be seen later, the same approximation
will not rule out the influence of an axial magnetic field component Bz .
We first turn to Eq. (9.133) which is multiplied by the factor
2
2
(m
r ) , to result in
r 2
r
1
2 2
Er '
1k
+
(iE )
(9.136)
m
r
r
in a first approximation. When inserting this relation into Eq. (9.134)
the latter is found to become identically satisfied for small . Finally,
Eq. (9.135) also becomes approximately satisfied for Ez ' 0 and small .
With the result (9.136) and Ez ' 0, the component iE can be used
as a generating function F which determines the total electric field
118
Bo Lehnert
G = R () eim ,
(9.137)
v = c (sin ) ,
(9.139)
Z +
ik
z
Ak e dk =
exp
+ ik0 z .
(9.140)
k0 z0
2z0
2
1
2 2 r0
Er = E0 1 k0
+
R ()(sin m)
,
m
m
where
#
"
2
z
+ ik0 z ,
E0 = a0 exp
2z0
(9.141)
(9.142)
a0 =
G0
,
k0 z 0
(9.143)
The
corresponding magnetic field becomes
k0
k0
m
1
B = i (iE ) ,
Er ,
Er +
(iE ) . (9.144)
r0
r0
119
z = k 2 2 E0 r0 1 2 + 1 R (sin m)
(9.145)
B
.
0
m
2
In the analysis which follows we choose a function
"
2 #
z
E0 (
z ) = a0 exp
(sin k0 z) ,
2z0
(9.146)
Z
Z 2Z
r d
q = 0 (div E) dV = 0 r0
E
z +
0
0
(9.147)
Z +
+m
(iE ) d
z d d = 0 .
This result holds even before carrying out the integration with respect to and , because q disappears on account of the antisymmetry
r and iE
. A disappearing total charge would also result from the
of E
obtained surface integral due to Gauss theorem, provided that the field
quantities vanish at the origin and at infinity. Since we shall later also
treat a function R which is divergent at the origin, however, the result
q = 0 will here be based on the antisymmetry just being mentioned.
The integrated magnetic moment further becomes
Z +Z 2 Z
2
M = 0
(div E) c (cos ) r dr d d
z =
0 c (cos ) r02
Z
Z
0
2Z
0
) d
(iE
z d d = 0
Er d
z +
(9.148)
+m
for the same symmetry reasons as those applying to the total charge.
In the local energy equation (3.20) we consider the first term of the
z is neright-hand member. With the present approximation where E
glected and C = c is small, the contribution from this term becomes
negligible as compared to the second term. Consequently the energy
120
Bo Lehnert
density can be written as given by the form (3.25). The main contributions to wf then originate from the components (Er , E , Br , B ).
Here the periodic dependence and the phase differences have to be taken
into account when performing an integration over a period of oscillation
and over the volume elements. With the field components given by
Eqs. (9.141)(9.144), the average local energy density of the packet can
be written as
1
2 )
r2 + E
w
= 0 (E
(9.149)
2
in the present approximation. From expressions (9.146) and the integral
2 Z +
Z +
2
G0
2
z=
E0 d
(sin k0 z)2 e2(z/2z0 ) d
z '
k
z
0 0
(9.150)
3/2
' G20 2
k0 z0 2
for a small line width, we then obtain an integrated total mass
Z
1
1
m= 2
(9.151)
w
dV = 2 A0 r02 Wm
c
c
with
z0
A0 = 5/2 0 G20 2 2 ,
(9.152)
k0 z0 8
)
2
Z (
2
Wm =
d .
(9.153)
+1 R +R
m
Here m = 0 and m 6= 0 for the options of convergent and divergent
forms at the origin = 0 of the function R.
Considering the momentum equation (3.16) we introduce the field
E0 = E + C B .
(9.154)
(9.155)
0 = 0 ,
E
(9.156)
z0
E
(9.157)
(cos ) .
=E
121
of Eq. (3.16) it is further observed that it includes the factors div E and
E which vary as sin m
and cos m
due to Eqs. (9.110)(9.112) and
(9.114). Integration of the axial force from = 0 to = 2 then yields
a vanishing net result. Only the electromagnetic momentum density
(3.19) thus remains to be considered for the wave packet in its entirety.
The related angular momentum density is then given by the conventional expression (9.76). In the present approximation it reduces to
k0 2
2 ), rE
r B
z
s ' 0 0, r
(Er + E
(9.158)
r | is proportional to |iE
| according to Eq. (9.114), and Er 1
since |B
iE
as shown later here when having specified the radial function R ().
The axial spin component sz is of particular interest to this study.
Its volume integral becomes
Z
1 k0
A0 r04 2 Ws ,
(9.159)
s sz dV = 3
m
c
where
2
Z
3
Ws =
+ 1 R d .
(9.160)
s
In analogy with the integral (9.153) the lower limits are here s = 0
and s 6= 0 for convergent and divergent forms of R. The result (9.159)
shows that there are two possible spin directions, depending upon the
sign of m.
N hc
,
0
1 2
Nh
A0 r04 2 Ws =
|m|
3 c0
2
(9.161)
(9.162)
122
Bo Lehnert
1/2
Wm
0
|m|
3/2
2r0 =
(9.163)
Ws
both for a dense beam and for an individual photon, under the conditions just being specified.
To proceed further with the analysis, the radial part R () of the
generating function and corresponding integrals have now to be specified. This concerns two cases, namely when R is convergent all over
space, and when R diverges at the origin, in analogy with the earlier
analysis of Sections 9.3.4 and 9.3.5. One option is to adopt a radial
part R which is finite at the origin = 0 and which vanishes at large .
These requirements are met by
r
R = e ,
=
(9.164)
,
r0
where 1. As in 9.3.4 the final result then becomes independent of
. The form (9.164) has a rather sharp maximum at = = , and
r = r0
(9.165)
0
|m|
3/2
(9.168)
which is independent of .
With the function (9.164) and expressions (9.141) and (9.142) the
relative magnitude of the electric field components further becomes
E
r | + 1 | 1,
(9.169)
iE
123
3 hc2 3
3
' 1.85 1015 4 [W/m2 ]
4
0
0
(9.170)
according to Eq. (9.107). With the data of the previous numerical example this results in a maximum energy flux p ' 2 W/m2 for the photon
fields not to overlap. In many cases of practical interest, the diameter
(9.168) could therefore apply to models of dense photon beams.
In the case of such beams it is finally observed that the energy density (9.149) determined by Eqs. (9.141), (9.142) and (9.164) leads to a
radial distribution of intensity which forms a ring-shaped region with
an annular radius of the order of the radius r given by Eq. (9.168).
The vanishing field strengths at the axis = 0 of such a ring-shaped
distribution thereby support the use of the approximate dispersion relation (9.22), because there are then negligible contributions to the field
intensity from regions near the axis.
Turning finally to the alternative of a radial part of the generating
function which is divergent at the axis = 0, the form
R () = e
(9.171)
124
Bo Lehnert
cr > 0 ,
G0 = cG ,
cG > 0 ,
(9.174)
> 0,
(9.175)
N hc
2+2
,
22 =
m
0
(9.176)
Nh
2 1 2+4
=
,
24
3
c0 |m
| s
2
(9.177)
A0
2
where A0 = 2G
2 . For finite values of mc and |s| it is thus required that
0
m = (2+2)/(22) ,
s = (2+4)/(24) .
(9.178)
(9.179)
0
.
|m|
3/2
(9.181)
125
With the function (9.171) and expressions (9.141) and (9.142) the
divergent case now gives
E
r | 1 + | 1
(9.182)
iE
for
E all
when 1. Also here, as in the case (9.164), we obtain
r 1. A corresponding individual photon or beam then becomes
iEz
|m
|m|
0
r c (cos )|
'
|k0 c (sin )|
2 r
(9.183)
|m|
5/2 2
.
(9.184)
= C0 = const,
(9.185)
C 0 0
,
|m|
3/2
(9.186)
|m|
5/2
.
C02
(9.187)
Fm0 =
0 =3 107 m,
2
6
and Fm0 = 10 , say, from which C0 = 10 and 2 r = 10 m. This value
126
Bo Lehnert
is much smaller than that obtained from the convergent model. It exceeds on the other hand atomic dimensions by a substantial factor, this
being due to the limit of validity of the approximation imposed on the
dispersion relation (9.22). The result (9.186) can in any case be taken as
an indication in the direction towards a strongly needle-shaped individual photon configuration. Such a geometry at atomic dimensions could
possibly become realizable in an exact analysis of the basic equations
without approximations, but a corresponding treatment remains to be
performed.
For dense light beams of a comparatively small cross-section, condition (9.187) appears to become rather well satisfied. With an example
of m
= 1, 0 = 3 107 m, Fm0 = 106 and C0 = 103 one would obtain
2 r = 104 m.
9.6.4
127
128
Bo Lehnert
Chapter 10
SUPERLUMINOSITY
The possible existence of objects traveling faster than the velocity c of
light, but at a limited speed, has a long history which can be traced
back to the early 1900s, as described in reviews by Recami100,101, Barut
et al.102, and Cardone and Mignani103 among others. The physics of
superluminal phenomena was regarded until the mid 1990s by most
physicists as a waste of time, or worse. First after a number of new
observations and experimental facts, mainly being found since 1991,
the interest in superluminal processes has been revived, even if these
processes are still subject to controversial discussions. This chapter is
devoted to a brief description of some main points in the research on
superluminosity.
10.1
Tachyon theory
The problem of faster-than-light particles was first reconsidered by Bilaniuk et al.104, and G. Feinberg later called these objects tachyons.
The Italian school headed by E. Recami generalized special relativity
to superluminal inertial frames100 and introduced the term extended
relativity. The latter concept was based on the two postulates of the
principle of relativity and of the homogeneity of spacetime and isotropy
of space. Thereby the vacuum speed c of light is generalized to become
a two-side limit, in the sense that one can approach it either from below
or from above. The two basic postulates imply that the metric tensor
is invariant for subluminal velocities, and also invariant except for its
sign in the case of superluminal velocities. In two dimensions the corresponding generalized Lorentz transformation by Recami and Mignani
then becomes100
x0 x
x00 = q
,
(10.1)
|1 2 |
x x0
,
x0 = q
|1 2 |
(10.2)
where = uc and < u < +. Superluminal Lorentz transformations in four dimensions are somewhat more complicated, because
130
Bo Lehnert
(10.3)
with Et , pt , and mt0 as the energy, momentum, and rest mass of the
tachyon100. Here m2t0 < 0 is interpreted as an imaginary mass.
The central concepts of tachyon theory can also be made to emerge
from the present theory. Thus, the condition C2 = c2 of Lorentz invariance given by the second of Eqs. (3.3) is not only satisfied by velocity
vectors of the form (9.1) in the subluminal case, but also by vectors of
the form
C = c (0, i sinh , cosh ) = c (0, C , Cz )
(10.4)
in a superluminal
case. For propagating normal modes varying as
exp i (t + kz) , Eq. (4.3) for an EMS-like tachyon mode then yields
the dispersion relation
= k Cz = k u ,
u = c (cosh )
(10.5)
(10.7)
This rest mass therefore becomes imaginary, in accordance with current tachyon theory.
Comparison between expressions (10.4) and (9.1) indicates that cos
has to be replaced by i(sinh ), and sin by cosh , to convert the earlier
deduced subluminal results of Chapter 9 into a corresponding axisymmetric tachyon superluminal wave-packet theory. Care is then necessary
in not confusing the imaginary unit used to indicate the phase in time
with the same unit applied in the velocity expression (10.4). Eqs. (9.85)
and (9.95) yield imaginary transverse tachyon diameters. The physical
interpretation of an imaginary photon rest mass and diameter is not
clear at this stage.
Chapter 10 Superluminosity
131
132
Bo Lehnert
Chapter 11
NONLOCALITY
Another class of electromagnetic phenomena beyond the concepts of
conventional theory is characterized by nonlocality, in the form of instantaneous long-range interaction. Also these phenomena have recently
attracted an increasing interest, and have become the object of extensive
and partly controversial discussions.
11.1
General questions
It has been pointed out by Dirac63 that, as long as we are dealing only
with transverse waves, Coulomb interaction cannot be included. There
must then also arise longitudinal interactions between pairs of charged
particles. Argyris and Ciubotariu114 further notify that the unquantized
longitudinal-scalar part of the electromagnetic field yields the Coulomb
potential, and that transverse photons transport energy whereas longitudinal (virtual) photons do not carry energy away. Thus, there is
direct interaction between a transverse photon and the gravitation field
134
Bo Lehnert
of a black hole, but not with a longitudinal photon. The Coulomb field
is therefore able to cross the event horizon of a black hole.
In their considerations on the field generated by a single moving
charged particle, Chubykalo and Smirnov-Rueda5355 have claimed the
Lienard-Wiechert potentials to be incomplete, by not being able to describe long-range instantaneous Coulomb interaction. This question is
still under discussion10.
A further analysis on long-range interaction has been performed in
which the Proca-type field equation is subdivided into two parts5355.
The first part manifests the instantaneous and longitudinal aspects of
the electromagnetic nature, as represented by functions f [R(t)] of an
implicit time dependence. For a single charge system this would lead
to the form R(t) = r rq (t), where r is a fixed vector from the point of
observation to the origin, and rq (t) is the position of the moving charge.
The implicit time dependence then implies that all explicit time derivatives are left out from the part of the basic equations which belongs
to the instantaneous interaction. The second part, given by functions
g (r, t) where the explicit time derivatives are retained, is connected with
propagating transverse EM waves. This proposed subdivision can at a
first sight be commented on, from two different points of view:
The charges and currents are the sources of the electromagnetic
field, and they emit signals which reach a field point after a certain transit time. When the time variation of the same sources
becomes slow enough for the transit time of an EM wave to be
very short, then the velocity of light would appear as infinite, and
the terms with explicit time derivatives in Eq. (2.1) can be neglected with good approximation. As a result of this, it would not
become possible to distinguish such wave propagation from a real
instantaneous interaction based on an implicit time dependence
where the explicit time derivatives are exactly equal to zero;
There are, however, a number of arguments that appear to support the existence of long-range instantaneous interaction as described in 11.1, as well as the interesting idea of subdividing the
Proca-type equation. Further supporting arguments are based on
the longitudinal wave concept itself7, and on observations of the
universe113. Instantaneous interaction, represented by longitudinal components, can thus be interpreted as a classical equivalent
of nonlocal quantum interaction53.
The proposed subdivision of the basic equations into long-range and
short-range parts can also be applied to the present extended field equa-
Chapter 11 Nonlocality
135
To an increasing extent
it has been realized that the electromagnetic
four-potential A = A, i
does not only serve the purpose of a mathec
136
Bo Lehnert
(11.1)
,
(11.2)
t
according to Eq. (3.9). Consequently the basic equations for the S wave
reduce to
= 0
C div +
(11.3)
t
since Eq. (3.7) becomes automatically satisfied.
E = ,
= +
Chapter 11 Nonlocality
11.4.2
137
To study the joint effects of the curl-free vector potential and the S wave,
a simple time-dependent case is now chosen where there is cylindrical
geometry of the Aharonov-Bohm type. The long straight coil has its
axis at r = 0 and is extended along the z direction of a cylindrical frame
(r, , z). The coil cross section has the radius R0 . In the interior of the
coil there is a time-dependent homogeneous magnetic field B(t) being
directed along z. The corresponding total magnetic flux becomes
(t) = R02 B(t) .
(11.4)
(11.6)
=
(t)
2
in analogy with an example on the Aharonov-Bohm effect treated by
Ryder21. Here it is observed that becomes a multivalued function of
in a non-simply connected vacuum region. Thereby the theorem by
Stokes has to be applied with care, but the line integral of A around a
circle in the direction still becomes nonzero. It is now possible to write
1 d
=
,
, 0
(11.7)
r 2r dt
and
r
.
(11.8)
div =
r r
r
To satisfy Eq. (11.3), the velocity vector C must have components
both in the r and the directions, i.e.
C = c (sin , cos , 0) .
Then Eq. (11.3) leads to the system
c (sin )
r
+
r
= 0,
r
r
t
r
1 d2
r
+
= 0.
c (cos )
r
r
2 dt2
(11.9)
(11.10)
(11.11)
138
Bo Lehnert
1
= f (r ct)
(11.12)
r
r
for propagation in the positive r direction at the speed c of light. The
corresponding wave equation (11.10) is similar to but not identical with
that of sound waves in a compressible medium117. The wave equations
for these two phenomena have the same form for plane waves only.
In spherical geometry117 there is a solution being similar to that of
Eq. (11.12), but with the factor r12 ahead of f (r ct).
When 6= 0 there is an entirely new situation. To find the solution
of Eqs. (11.10)(11.11) an ansatz
r
r
=cf t
(11.13)
r
c
E=
sin = 1 ,
0 < 1,
(11.15)
there is a velocity (11.9) being mainly oriented along the radial direction
and circulating slowly around the z axis. Eqs. (11.10) and (11.11) then
become
cf 0 = 0 ,
(11.16)
2
1
d
c (2)1/2 f 0 =
(11.17)
.
2 dt2
To satisfy this system for all (r, t), and for finite values of f 0 and
d
, the following procedure has to be adopted. For a constant and
dt2
finite value of c it is readily seen that there are no solutions. Therefore
the velocity c has to be ascribed a new unconventional interpretation.
In order to prevent the left-hand member of Eq. (11.17) from vanishing
with , the constant c has now to tend to infinity, such as to satisfy the
condition
(11.18)
c 1/2 = c0 ,
2
Chapter 11 Nonlocality
139
1
d2
r
= f0 t
(11.20)
f 0 (t) ,
(c )
2
c
2 2 c0 dt
where all members become functions of t only. The characteristic length
Lc = |f (f /r)1 | and the characteristic time tc = | f (f /t)1 | are
then related through Ltc = 1c 0.
c
Consequently, this analysis indicates that the underlying equations
can only be satisfied in the limit of instantaneous interaction at a distance. Then the Lorentz condition (2.4) reduces to that of the Coulomb
gauge, and the divergence of A vanishes in agreement with Eq. (11.6).
The S wave signal of the curl-free and time-dependent magnetic coil
field is thus expected to arrive instantaneously at a point far from the
origin. The problem of detecting such an S-wave is, however, by no
means simple, nor is it even clear at this stage if such a signal can exist.
As obtained from an analysis by Angelidis118, the concept of instantaneous interaction does not become reconcilable with special relativity
which is based on a finite velocity c of propagation. The result (11.20)
can therefore also be understood as an example and a confirmation of
such an interpretation.
11.5
140
Bo Lehnert
Chapter 12
General conclusions
The main results which are specific to the present theoretical approach
can now be summarized. These are expected to contribute to an increasing understanding of a number of fundamental physical phenomena, and they also predict new features of the electromagnetic field to
exist. It should be noticed here that the recently presented theories
by Kaivarainen122 and Tauber123 have some parts in common with this
approach, such as the concept of a nonzero electric field divergence in
the vacuum state, models of the leptons with an internal vortex-like
structure, and a small but nonzero photon rest mass.
Maxwells equations with a vanishing electric field divergence in
the vacuum have been used as a guideline and basis in the development of quantum electrodynamics (QED)33,73,74. Therefore QED is expected also to be subject to the typical shortcomings of conventional
electromagnetic theory. The revised electromagnetics described here,
with its nonzero electric field divergence, provides a way of eliminating
these shortcomings in a first step of extended quantum electrodynamics (EQED). The present theory is thus based on the hypothesis of
a vacuum state which does not merely consist of empty space but can
become electrically polarized, to give rise to a local electric space charge
density and a related nonzero electric field divergence. On account of
this, and of the condition on Lorentz invariance, an additional spacecharge current density arises as a matter of necessity. It occurs along
with the conventional displacement current in the resulting extended
field equations. Maxwells equations thus become a special class of this
approach. The nonzero electric field divergence introduces an additional
degree of freedom, and this leads to a number of new electromagnetic
properties. From the analysis performed in the previous chapters, it is
thereby seen that the present theory has features and leads to results
which in several respects become directly or indirectly supported by
experimental observations.
So far quantum field theories have all incorporated Lorentz invariance in their basic structure, but its breaking has recently become subject of several investigations124. Whatever the true origin of a possible
142
Bo Lehnert
Lorentz breaking may be, the fact that it has not yet been observed
means it must be small at the energy scales corresponding to known
standard-model physics.
The present theory also has the important feature of being gauge
invariant, due to the form of the space-charge current density. This
invariance does not necessarily hold for other forms of the four-current
in a Proca-type field equation.
The present theory leads to some results of a general character, of
which the following should be mentioned:
The extended equations make it possible for electromagnetic
steady states to exist in the vacuum;
New types of wave modes arise, such as a longitudinal purely electric space-charge (S) wave, and a nontransverse electromagnetic
space-charge (EMS) wave.
12.2
143
144
Bo Lehnert
Long mean free paths are predicted for neutrinos in their interaction with solid matter, possibly being in agreement with observed data.
12.3
145
Bibliography
1. Feynman R. P. Lectures on Physics: Mainly Electromagnetism and Matter. Addison-Wesley, Reading, MA, 1964, p. 281.
2. Lakhtakia A. (Ed.) Essays on the Formal Aspects of Electromagnetic
Theory. World Scientific, Singapore, 1993.
3. Sachs M. Relativity in Our Time. Taylor and Francis, 1993; in Modern Nonlinear Optics, Part 1, Second Edition, Advances in Chemical
Physics, vol. 119, Edited by M. W. Evans, I. Prigogine, and S. A. Rice,
John Wiley and Sons, Inc., New York, 2001, p. 667.
4. Evans M. and Vigier J.-P. The Enigmatic Photon. Kluwer, Dordrecht,
vol. 1, 1994 and vol. 2, 1995.
5. Barett T. and Grimes D. M. (Eds.) Advanced Electromagnetics. World
Scientific, Singapore, 1995.
6. Evans M. W., Vigier J.-P., Roy S., and Jeffers S. The Enigmatic Photon.
Kluwer, Dordrecht, vol. 3, 1996.
7. Evans M. W., Vigier J.-P., Roy S., and Hunter G. (Eds.) The Enigmatic
Photon. Kluwer, Dordrecht, vol. 4, 1998, pp. 1 and 33; Evans M. W.
Found.Phys. Lett., 2003, v. 16, 367.
8. Hunter G., Jeffers S., and Vigier J-P. (Eds.) Causality and Locality in
Modern Physics. Kluwer, Dordrecht, 1998.
9. Lehnert B. and Roy S. Extended Electromagnetic Theory. World Scientific, Singapore, 1998.
10. Dvoeglazov V. V. (Ed.) Contemporary Fundamental Physics. Nova Science Publishers, Huntington, New York, 2000.
11. Evans M. W., Prigogine I., and Rice S. A. (Eds.) Modern Nonlinear Optics. Parts 13, Second Edition, Advances in Chemical Physics, vol. 119,
John Wiley and Sons, Inc., New York, 2001, p. 250.
12. Amoroso R. L., Hunter G., Kafatos M., and Vigier J.-P. (Eds.) Gravitation and Cosmology: From the Hubble Radius to the Planck Scale.
Kluwer, Dordrecht, 2002.
13. Lehnert B. Report TRITA-EPP-79-13. Electron and Plasma Physics,
Royal Institute of Technology, Stockholm, 1979.
14. Lehnert B. Spec. Sci. Technol., 1986, vol. 9, 177; 1988, vol. 11, 49; 1994,
vol. 17, 259 and 267.
15. Lehnert B. Optik, 1995, vol. 99, 113.
16. Lehnert B. Physica Scripta, 1999, vol. T82, 89.
17. Lehnert B. In: Contemporary Fundamental Physics, V. V. Dvoeglazov
(Ed.), Nova Science Publ., Huntington, New York, 2000, vol. 2, p. 3.
Bibliography
147
18. Lehnert B. In: Modern Nonlinear Optics, Part 2, Second Edition, Advances in Chemical Physics, vol. 119, Edited by M. W. Evans, I. Prigogine, and S. A. Rice, John Wiley and Sons, Inc., New York, 2001, p. 1.
19. Lehnert B. Physica Scripta, 2002, vol. 66, 105; 2005, vol. 72, 359; 2006,
vol. 74, 139; Progress in Physics, 2006, vol. 2, 78; 2006, vol. 3, 43; 2007,
vol. 1, 27.
20. Lehnert B. In: Gravitation and Cosmology: From the Hubble Radius to
the Planck Scale, Edited by R. L. Amoroso, G. Hunter, M. Kafatos, and
J.-P. Vigier, Kluwer, Dordrecht, 2002, p. 125.
21. Ryder L. H. Quantum Field Theory. Cambridge Univ.Press, 1996,
Ch. 2.8, Ch. 10; p. 406; Ch. 1; Ch. 3.4; also Edition of 1987.
22. Tsuchiya Y., Inuzuka E., Kurono T., and Hosoda M. Adv. Electron
Phys., 1985, vol.64A, 21.
23. Heitler W. The Quantum Theory of Radiation. Third Edition, Clarendon Press, Oxford, 1954, p. 401; p. 409; p. 58.
24. Stratton J. A. Electromagnetic Theory. First Edition, McGraw-Hill
Book Comp., Inc., New York and London, 1941, Ch. VI, Sec. 6.7;
Sec. 1.10, Ch. II and Sec. 1.23, 5.15.2 and 9.49.8; Sec. 2.5; Sec. 2.19;
Sec. 2.8 and 2.14; Sec. 8.5 and 3.9.
25. Jackson J. D. Classical Electrodynamics. John Wiley and Sons, Inc.,
New York-London-Sydney, 1962, Ch. 6, p. 201, Ch. 17, p. 589.
26. Thomson J. J. Nature, 1936, vol. 137, 232.
27. Hunter G. and Wadlinger R. L. P. Physics Essays, 1989, vol. 2, 154; with
an experimental report by F. Engler.
28. Donev S. In: Photon: Old Problems in Light of New Ideas, Edited
by V. V. Dvoeglazov, Nova Science Publishers, Huntington, New York,
2000, p. 32.
29. H
utt R. Optik, 1987, vol. 78, 12.
30. Mazet A., Imbert C., and Huard S. C. R. Acad. Sci., Ser. B, 1971,
vol. 273, 592.
31. de Broglie L. and Vigier J.-P. Phys. Rev. Lett., 1972, vol. 28, 1001.
32. Vigier J.-P. Physics Letters A, 1997, vol. 234, 75.
33. Schiff L. I. Quantum Mechanics. McGraw-Hill Book Comp., Inc., New
York-Toronto-London, 1949, Ch. IV; Ch. I; Ch. XIV.
34. Casimir H. B. G. Proc. K. Ned. Akad. Wet., 1948, vol. 51, 793.
35. Lamoreaux S. K. Phys. Rev. Letters, 1997, vol. 78, 5.
36. Richtmyer F. K. and Kennard E. H. Introduction to Modern Physics.
McGraw-Hill Book Comp., Inc., New York and London, 1947, p. 600.
37. Dirac P. A. M. Proc. Roy. Soc., 1928, vol. 117, 610 and 1928, vol. 118,
351.
38. Morse P. M. and Feshbach H. Methods of Theoretical Physics. McGrawHill Book Comp., Inc., New York 1953, Part I, p. 260 and 331; Part II,
p. 1320.
148
Bo Lehnert
Bibliography
149
150
Bo Lehnert
Bibliography
151
107. Nimtz G., Enders A., and Spieker H. In: Wave and Particle in Light
and Matter, A. van der Merwe and A. Garuccio (Eds.), Proc. Trani
Workshop (Italy, Sept. 1992), Plenum, New York, 1992.
108. Saari P. and Reivelt K. Phys. Rev. Lett., 1997, vol. 79, 4135.
109. Walker W. D. In: Gravitation and Cosmology: From the Hubble Radius
to the Planck Scale, Edited by R. L. Amoroso, G. Hunter, M. Kafatos,
and J.-P. Vigier, Kluwer, Dordrecht, 2002, p. 189.
110. Cufaro-Petroni N., Dewdney C., Holland P. R., Kyprianidis A., and
Vigier J.-P. Foundations of Physics, 1987, vol. 17, 759.
111. Aharonov Y. and Bohm D. Phys. Rev., 1959, vol. 115, 485.
112. Cufaro-Petroni N., Droz-Vincent P., and Vigier J.-P. Lettre al Nuovo
Cimento, 1981, vol. 31, 415.
113. Pope N. V. In: Causality and Locality in Modern Physics, Edited by
G. Hunter, S. Jeffers, and J.-P. Vigier, Kluwer, Dordrecht, 1998, p. 187.
114. Argyris J. and Ciubotariu C. In: Causality and Locality in Modern
Physics, Edited by G. Hunter, S. Jeffers, and J.-P. Vigier, Kluwer, Dordrecht, 1998, p. 143.
115. Kafatos M. In: Causality and Locality in Modern Physics, Edited by
G. Hunter, S. Jeffers, and J.-P. Vigier, Kluwer, Dordrecht, 1998, p. 29.
116. Chambers R. G. Phys. Rev. Letters, 1960, vol. 5, 3.
117. Lehnert B. Found. Phys. Lett., 2002, vol. 15, 95.
118. Angelidis T. Private communication, 2003.
119. Puthoff H. E. Phys. Rev. D, 1987, vol. 35, 3266; Phys. Rev. A, 1989,
vol. 39, 2333.
120. Haisch B., Rueda A., and Puthoff H. E. Phys. Rev. A, 1994, vol. 49, 678.
121. Bearden T. E. Energy from the Vacuum. Cheniere Press, Santa Barbara,
California, 2002.
122. Kaivarainen A. Bivacuum Mediated Electromagnetic and Gravitational
Interactions. In: Frontiers in Quantum Physics Research, Edited by
F. Columbus and V. Krasnoholvets, Nova Science Publishers, Inc., 2004,
pp. 83128.
123. Tauber H. Physics Essays, 2003, vol. 16, 200.
124. Pospelov M. and Romalis M. Lorentz Invariance of Trial. Physics Today,
July 2004, p. 40.
Author Index
Afshar S. S. 105, 150
Aharonov Y. 133, 151
Amoroso R. L. 1, 11, 20, 146
Anastasovski P. 19, 148
Angelidis T. 139, 151
Argyris J. 133, 135, 139, 151
Arshansky R. I. 24, 149
Barrett T. 11, 146
Bartlett D. F. 17, 148
Barut O. A. 129, 150
Bass L. 16, 19, 89, 100, 148
Battersby S. 112, 126, 150
Bearden T. 19, 140, 148, 151
Behroozi C. H. 103, 150
Bethe H. A. 76, 150
Bilaniuk O. M. 129, 150
Bohm D. 133, 151
Borb
ely T. 19, 108, 148
Broglie L. de 14, 16, 89, 95, 100, 102104,
108, 147
Bruce S. 24, 149
Cardone F. 129, 131, 150
Casimir H. B. G. 15, 147
Chambers R. G. 136, 151
Chubykalo A. E. 18, 27, 133, 134, 148
Ciubotariu C. 19, 133, 135, 139, 148, 151
Close F. 76, 150
Coffey W. I. 19, 148
Cohen E. R. 55, 149
Corle T. R. 17, 148
Crowell L. B. 19, 148
Cufaro-Petroni N. 133, 136, 151
Deshpande V. K. 129, 150
Dewdney C. 133, 136, 151
Dirac P. A. M. 16, 20, 23, 24, 45, 133,
147, 149
Ditchburn R. W. 111, 127, 150
Donev S. 13, 31, 147
Droz-Vincent P. 133, 151
Dutton Z. 103, 150
Dvoeglazov V. V. 11, 24, 103, 134, 146,
149
Einstein A. 16, 20, 38, 45, 89, 100, 104,
131, 133, 148, 150
Author Index
Kennard E. H. 15, 147
Knoesel E. 105, 150
Kurono T. 12, 104, 128, 147
Kyprianidis A. 133, 151
Labounsky A. 19, 148
Lakhtakia A. 11, 146
Lamoreaux S. K. 15, 147
Land M. C. 24, 149
Lehnert B. 11, 13, 17, 19, 33, 46, 55, 56,
63, 71, 72, 9598, 100, 102,
111, 128, 138, 146151
Leighton R. B. 25, 149
Maccarone G. D. 129, 150
Mair A. 112, 150
Marten M. 76, 150
Mathews P. T. 19, 148
Mazet A. 14, 147
McDonald K. F. 105, 150
M
ez
aros M. 19, 108, 148
Michelson A. A. 89, 101, 123, 144
Mignani R. 129, 131, 150
Moln
ar P. R. 19, 108, 148
Morley E. W. 89, 101, 123, 144
Morse P. M. 16, 95, 147
M
unera H. A. 20, 148
Nambu Y. 44, 149
Nimtz G. 131, 151
Ohmura T. 9, 20, 148
Olkhovsky V. S. 131, 132, 150
Particle Data Group 20, 149
Planck M. 19, 45, 108
Polyakov A. M. 20
Pope N. V. 133, 151
Pospelov M. 143, 151
Prigogine I. 11, 146
Puthoff H. E. 140, 151
Recami E. 129, 131, 132, 150
Reivelt K. 132, 151
Rice S. A. 11, 146
Richtmyer F. K. 15, 147
Romalis M. 143, 151
Roy M. 18, 148
Roy S. 11, 1820, 27, 108, 135, 146, 148
Rueda A. 18, 140, 148, 151
Ryder L. H. 12, 20, 45, 55, 106, 137, 147
153
Saari P. 132, 151
Sachs M. 11, 20, 146, 149
Sarfatti J. 20, 149
Scheffel J. 46, 55, 56, 67, 71, 150
Schiff L. 15, 31, 104, 141, 144, 147
Schr
odinger E. 16, 19, 89, 100, 148
Schwinger J. 45, 63, 149
Sherwin C. W. 16, 148
Shnerb N. 24, 149
Sinha M. 18, 148
Smirnov-Rueda R. 18, 27, 133, 134, 148
Snellman H. 76, 150
Soad D. 24, 149
Spiegel M. R. 94, 150
Spieker H. 131, 151
Stratton J. A. 13, 18, 2729, 31, 39, 68,
8183, 86, 115, 147
Sudarshan E. C. G. 129, 150
Sutton G. 76, 150
Taylor B. N. 55, 149
Tauber H. 141, 151
Thomson J. J. Sir 13, 86, 147
Tsuchiya Y. 12, 104, 128, 147
Vaziri A. 112, 150
Vestergaard Hau L. 103, 150
Vigier J.-P. 11, 12, 14, 1620, 27, 89,
100103, 108, 133, 135, 136,
146148, 151
Wadlinger R. L. P. 13, 86, 98, 110, 147
Walker W. D. 132, 151
Weber J.64, 150
Weihs G. 112, 150
Zeilinger A. 112, 150
Subject Index
Aharonov-Bohm effect 133
Angular momentum 38, 95, 111
electron 44
neutrino 44
photon 12
Antimatter 42
Axisymmetric EMS modes 85, 88
angular momentum 95
axial magnetic field 89
beams 96
charge 93
dispersion relation 88
effective photon diameter 97
field components 88
generating function 88
helical geometry 89
Lorentz transformation 90
magnetic moment 93
phase and group velocities 88
radial polarization 88, 89
rest frame 90
rest mass 95, 100
total mass 94
two-slit experiments 104
wave packets 91
wave-particle concept 103
Beam of photons 109
Bosons in general 106
Casimir effect 15
Characteristic radius 36, 88
Compton wavelength 51
Convection displacement current 18
Conventional wave modes 86, 111, 114
Cylindrical wave modes 84
Field equations 22
Fine-structure constant 45
Force balance 26, 31, 68, 95, 120
Four-current 16, 21
Four-dimensional representation 15
current density 16, 21
field equations 15
potentials 15
Fresnel laws 13, 81
Subject Index
polar part 39, 41
radial part 39, 40
screw-shaped wave 116, 117
separable form 39
steady states 36
symmetry properties 39
Goos-H
anchen effect 13, 102
Group velocity 86, 88, 116, 117
Helical field geometry 89, 101, 144
Evans-Vigier photon model 89
Hertz theory 18
Higgs particle 106
Hubble redshift 18
Instantaneous interaction:
see nonlocality
Integrated field quantities 38
steady states 38, 73, 74
wave modes 93, 119
Lagrangian density 24
Lienard-Wiechert potentials 132, 134
Light beam 109, 128
angular momentum 111
Longitudinal magnetic field 17, 19, 89
Longitudinal waves 108, 136, 139
Long-range interaction:
see nonlocality
Lorentz condition 16
Lorentz invariance 21, 102
electron radius 55
Magnetic field divergence 23
Magnetic flux 51
Magnetic island 52, 53
Magnetic moment 45, 46
Dirac theory 45
electron model 45
Land
e factoe 45, 71
particle-shaped states 45
Schwinger-Feynmann correction 45
wave modes 93, 119
Magnetic monopoles 19, 23
Magnetic permeability 23
Magnetic vector potential 16
curl-free 24, 135
Michelson-Morley experiment 89, 101
Momentum equation 26
Near-field approximation 48
155
Needle radiation 12, 99, 104, 107, 144
Neutrino 72
mass 75
radius 76
Neutrino model 72
angular momentum 73, 74
effective radius 73
features 76
force balance 77
free path 76
rest mass 73, 74
Neutrino oscillations 102, 106
Nicols prism 128
Nonlocality 133
curl-free vector potential 136
electromagnetic 133
gravitational 135
propagation of energy 139
S wave 136
Pair formation 15
Phase velocity 86, 88, 117
Photoelectric effect 12, 104, 107, 144
needle radiation 2, 84, 88
Photon gas 19, 108
Plancks radiation law 19, 108
thermodynamics 108
Photon models 13, 84, 100, 112
other bosons 106
Photon oscillations 106
Photon radius 97, 99, 122, 123
Photon rest mass 16, 19, 89, 95, 100, 102
Planck length 67
Plane waves 78
decay 81
dispersion relations 8, 78
total reflection 80
wave types 78
Point-charge-like state 47, 142
Polarization of charge 15, 21
Polarization of field 127
Poynting theorem 26, 28
Poynting vector 26, 28, 144
Proca-type equations 15
Quantization 25
field equations 26
Quantized momentum operator 96
Quantum conditions 25, 44, 54
angular momentum 44, 54, 74, 95,
96, 99, 121, 124
156
Bo Lehnert
axisymmetric 84
cylindrical 84
plane 78
screw-shaped 112
Wave packets 91, 118
Wave-particle dualism 103, 144
Bo Lehnert. A Revised Electromagnetic Theory with Fundamental Applications. Svenska fysikarkivet, 2008, 158 pages.
ISBN 978-91-85917-00-6
Summary: There are important areas within which the conventional electromagnetic theory of Maxwells equations and its combination with quantum
mechanics does not provide fully adequate descriptions of physical reality. As
earlier pointed out by Feynman, these difficulties are not removed by and are
not directly associated with quantum mechanics. Instead the analysis has
to become modified in the form of revised quantum electrodynamics, for instance as described in this book by a Lorentz and gauge invariant theory. The
latter is based on a nonzero electric charge density and electric field divergence in the vacuum state, as supported by the quantum mechanical vacuum
fluctuations of the zero-point energy. This theory leads to new solutions of
a number of fundamental problems, with their applications to leptons and
photon physics. They include a model of the electron with its point-chargelike nature, the associated self-energy, the radial force balance in presence
of its self-charge, and the quantized minimum value of the free elementary
charge. Further there are applications on the individual photon and on light
beams, in respect to the angular momentum, the spatially limited geometry
with an associated needle-like radiation, and the wave-particle nature in the
photoelectric effect and in two-slit experiments.
Hannes Alfvens
department. Lehnert was a guest researcher at the department of Professor Subrahmanyan
Chandrasekhar, Yerkes Observatory,
USA, 19531954, and received his
PhD (Tekn. Dr.) in electrodynamics at
the Royal Institute of Technology in
1955. He became full professor (personal chair) at the Swedish Atomic
Research Council in 1968, and was
a member of the International Fusion Research Council of the International Atomic Energy Agency in Vienna, 19701991. During 19801990
he served as head of the Swedish Fusion Research Unit being associated
with the Euratom Research Programme in Brussels. Lehnerts main fields of interest have been magnetohydrodynamics, cosmical plasma physics, controlled
thermonuclear fusion, and electromagnetic field theory, as reported in more
than 200 publications in international journals. Professor Lehnert is a member
of the Royal Swedish Academies of Sciences and Engineering Sciences, and
of the Electromagnetics Academy, Cambridge, Mass., USA.
1950, samt
genjorsexamen
vid Kungl. Tekniska Hogskolan
i Stockholm ar
var gastforskare
vid Professor Subrahmanyan Chandrasekhars institution vid
sin Teknologie DoktorsexYerkes Observatory, USA, 19531954, och erholl
for
Atomforskning ar
1968, och var
(personlig forskartjanst)
vid Statens Rad
ledamot av International Fusion Research Council vid International Atomic
Kungl. Ingenjorsvetenskapsakademien,
samt av Electromagnetics Academy, Cambridge, Mass., USA.